INFORMATION TO USERS
This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted.
The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction.
1.The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity.
2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame.
3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete.
4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced.
5. PLEASE NOTE: Some pages may have indistinct print. Filmed as received.
University Microfilms International 300 North Zeeb Road Ann Arbor, Michigan 48106 USA St. John's Road, Tyler's Green High Wycombe, Bucks, England HP10 8HR 77-2341 BALASUBRAMANIAN, Kattalampat Narayanaswamy, 1943- VISUAL FIELD OF DRIVERS AS A FUNCTION OF FOVEAL TASK DEMANDS AND TARGET VISIBILITY IN A SIMULATED HIGHWAY ENVIRONMENT. The Ohio State University, Ph.D., 1976 Engineering, industrial
Xerox University Microfilms,Ann Arbor, Michigan 48106 VISUAL FIELD OF DRIVERS AS A FUNCTION OF
FOVEAL TASK DEMANDS AND TARGET VISIBILITY
IN A SIMULATED HIGHWAY ENVIRONMENT
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University
by
K. N. Balasubramanian, B. E., M. Sc., M. S.
* * * *
Hie Ohio State University 1976
Reading Committee: Approved by
Dr. Thomas H. Rockwell Dr. Glenn A. Fiy Dr. John B. Neuhardt A dviser Dr. George L. Smith, Jr, 4 Department of Industrial and Systems Engineering ACKNOWLED GMENTS
This research was accomplished as part of the aims of the Ohio State University Research Foundation project RF 4224 Al, titled, ’’Utility of Periph eral Vision to Motor Vehicle Drivers”. The project was funded by the National Highway Traffic Safety Administration, Washington, D. C.
I would like to thank all the faculty and staff of the Department of Indus trial and Systems Engineering and the Systems Research Group for all the help extended to me during the execution of this research and during my studyperiod.
I am most grateful to Dr. Thomas H. Rockwell who served as advisor for this research and guided me in all respects throughout my Ph.D. program. Without his patient guidance and constant encouragement, my research and studies would not have been successfully completed.
I wish to express my special thanks and appreciation to Dr. Glenn A. Fry for spending many hours with me in the laboratory guiding me through every detail regarding visual and optical problems of the research.
I acknowledge with sincere thanks the comments and advise received from Drn John B. Neuhardt, Dr. George L. Smith, Jr. and Dr. Robert L. Wick, Jr. during this research.
I take this opportunity to thank the Ohio Department of Transportation for loaning me the photometers used in the luminance measurements, and specially Mr0 Wally Richardson for guiding me in the use of the photometers.
Thanks are due to Messrs. Clarence James, Larry Tracewell, and Leonard C. Samuelson of the Systems Research Group for their help in setting up the hardware and software in the laboratory. I also appreciate the assistance provided by Mr. John Friend and Mr. Thomas Snider of the Cinema, and Pho tography department during filming and editing. I extend my appreciation to Mr. Dave Selby and Mr. Dave Brickner of the Teaching Aids Laboratory for their help in the modification of 16 mm projectors.
ii I would like to express my gratitide to all my colleagues in the Systems Research Group, specially to Messrs. Joe Stafford, Omar Sawaf, and Tom Kretovics who had readily helped me during different phases of this research.
Thanks to Mrs. Dawna Kiesling who patiently typed both the draft and the final report, to Miss Cindy Bucher who typed portions of the final report, and to Mrs. Lois Graber who gave the final touch to the organization of the report.
Finally, I would like to express my love and gratitide to my wife Meera and son Babu for their patience and understanding during our stay in Columbus.
iii VITA
July 10, 1943 • • Bom - Kattalampat, Madras, India
1965 . . .
1965 - 1968 • • Technical Teacher Trainee College of Engineering, Madras - 25
1968 . . .
1968 - 1971 • • Lecturer, Department of Mechanical Engineering, College of Engineering, Madras - 25, India
1972 - 1973 « • Teaching Assistant, Department of Industrial and Systems Engineering, Ohio University, Athens, Ohio
1973 . . .
1973 - 1976 • • Research Associate, Systems Research Group, Department of Industrial and Systems Engineering, The Ohio State University, Columbus, Ohio
PUBLICATIONS
"A Theoretical and Experimental Investigation of Automobile Path Deviations When Steering with No Visual Input, " Ohio University, M. S. thesis, 1973.
"A Theoretical and Experimental Investigation of Automobile Path Deviations When Driver Steers with No Visual Input, " Transportation Research Record, 520, pp. 25-37, 1974.
"The Interactive Effects of Carbon Monoxide and Alcohol on Driving Skills, " (with T. H. Rockwell and F. W. Weir), Final Report, RF 3332, Systems Research Group, Research Foundation, The Ohio State University, 1975.
"Effects of Noxious Gases on Driver Performance, " (with T. H. Rockwell and R. L. Wick, J r.), Final Report, RF 3734, Systems Research Group, Research Foundation; The Ohio State University, 1974.
"Carbon Monoxide Effects on Highway Driving Performance: An Investigation of the Effects of 12 Percent COHb on the Nighttime Performance of Young and Aged
iv Drivers, " (with T, H, Rockwell), paper presented at the American Association for Automotive Medicine, Nineteenth Annual Conference, San Diego, Calif,, November 20-22, 1975.
’’Evaluation of Illumination Designs for Accident Reduction at High Nighttime- Accident Highway Sites, ” (with T. H. Rockwell and J. C. Hungerford), Final Report, EES 434, Systems Research Group, Department of Industrial and Systems Engineering, The Ohio State University, June 1976.
FIELDS OF STUDY
Major Field: Industrial and Systems Engineering
Human Factors Engineering. Professor Thomas H. Rockwell
Applied Statistics. Professor John B. Neuhardt
Decision Analysis. Professor William T. Morris
v TABLE OF CONTENTS
ACKNOWLEDGMENTS...... ii
VITA ...... iv
LIST OF TA B LES ...... vii
LIST OF FIGURES ...... ix
Chapter
I, THE ROLE OF PERIPHERAL VISION IN HUMAN VISUAL P E R F O R M A N C E ...... 1
n. ROLE OF PERIPHERAL VISION IN DRIVING • • • • 45
HI. THE RESEARCH PROBLEM AND METHODOLOGY . . 71
IV. LABORATORY SET UP AND INSTRUMENTATION . . 77
V. RESEARCH STUDIES: PURPOSE, PROTOCOL AND RESULTS »«**•••••••••• 109 VI. SUMMARY OF RESULTS AND CONCLUSIONS . . . 169
APPENDIX
A. ANOVA TABLES AND RESIDUAL PL O T S ...... 183
BIBLIOGRAPHY...... 221 LIST OF TABLES
Table Page 1 Factors Affecting Visual Perception through Peripheral Vision ...... * ...... 6
2 Area of the Entrance Pupil in Relative Values for Several Diameters d and Eccentricities r \ (after LeGrand, 1967) . . . 11
3 Astigmatism of the Peripheral linage (after LeGrand, 1967) . . 12
4 Cones in Retinal A reas ...... 13
5 Summary of Literature on the Effect of Foveal Task on Performance in the Visual Periphery ...... 23-27
6 Characteristics of Eye M ovements ...... 42
7 Foveal Task and Relevant Peripheral Targets for Some Common Driving Activities ...... 48
8 Resolution Angles in the 32° Binocular Visual Field of Drivers with Normal Vision (20/20)...... 56
9 Angular Sizes of Objects with a Clarity Index of 4. 0 at Different Eccentricities (Ford Study, 1973) ...... 57
10 Visual Field for the Left E y e ...... 59
11 Head Movements - Number and Distribution for Lane Change Maneuver (Left and R ig h t) ...... 61
12 lype of Motion Perception by Eye Position and Motion of Objects and Environment ...... 64
13 Visual Angles Subtended by Vehicles in the Adjacent Lanes . . . 82
14 Target Visibility for Different Wattages of Slide Projector . . 97
15 Data Absolute Contrast of Highway Vehicles by Bhise (1976) . . 97
v i l List of Tables Continued
16 Vehicle Contrasts from Luminances Measured in the Field . . 98
17 Stimulus Uncertainty Associated with Different Number of Alternatives in the Foveal T a s k ...... 101
18 Normal Visual Field Data in the Horizontal Meridan in Visual D egrees ...... 114
19 Data on Simple Reaction Tim e...... 117
20 Sample Output from the Analysis of the Data on Foveal Task Performance ...... 121
21 Foveal Task Performance Critical "On Time" by Subjects . . 122
22 Vehicle Image Size and Visual Angle subtended at the Subject's E y e ...... 125
23 Threshold Visibility Data for Three Different Conditions of Target Presentation in the Laboratory for Ten Subjects . . . 127
24 Peripheral Detection Angle (Stages I & II, Pilot Experiment). . 130
25 Target Size and the Visibility Details for the Four Slides used in Stage IV S t u d y ...... 142
26 Foveal Task Performance Data with and Without Peripheral Task for Subjects SOI and S 0 2 ...... 149
27 Mean and Standard Deviation of Response Times in a Foveal Task (in the presence of peripheral task) ...... 158
28 Matrix of Product Moment Correlation Coefficients between Subject Descriptors, Baseline Data, Foveal and Peripheral Task Performance M easures ...... 160
29 Correlation Matrix for PDA, Target Speed and Foveal Task E r r o r s ...... 164
30 Summary of Results from Pilot Experiments ...... 174
31 Summary of Results from Major E xperim ents ...... 175
viii LIST OF FIGURES
Figure Page
1 Model of Visual Perception ...... 3
2 Factors Affecting Visual Perception through Peripheral V isio n ...... 5
3 Three Basic Dimensions of Visual Response Criteria • • • 9
4 Simplified Diagram of a Cross Section of the Human Eye . • 10
5 The Limit of Peripheral Vision (after Pirenne, 1967) . . . 10
6 Variation of Transmission through Optical Media as a Function of Target Eccentricity (after Weale, 1956) . . . 11
7 Top View of the Left Eye, and the Corresponding Densi- 14 ties of Rods and Cones across the R e tin a ......
8 Dark Adaptation C u rv e ...... 16
9 Directional Sensitivity of C o n e s ...... 16
10 The Spectral Sensitivity Curves for Rods and Cones . . . 17
1 1 Total Stimulus Energy ( A L x T) at Threshold as a Function of Exposure Duration for Various Background Luminances (From Committee on Vision, National Re search Council, 1 9 7 5 ) ...... 19
12 Contrast Threshold as a Function of Retinal Locus for Different Background Luminances ...... 20
13 Data on Movement Thresholds ...... 32
14 Gould’s Model of the Ongoing Processes during an Individual Eye Fixation in a Visual Search T a sk ...... 43
ix List of Figures Continued
15 Cumulative Percent of Trials in which Target Vehicle was Detected as a Function of Target Eccentricity Angle. . . . 54
16 Schematic Diagram of the Laboratory Set Up for Screens and Projectors ...... 79
17 Diagram Showing the Relationship of Scene Image Size to Slide Image Size...... 84
18 Angular Speed of Vehicles Passing in the Left Adjacent Lane at Different Angular Locations for Different Constant .^Relative Velocities ...... * 85
19 Angular Speed of Vehicles Passing in the Left Second Lane at Different Angular Locations for Different Constant Rela tive V elocities ...... 8 6
20 Schematic Diagram of the Set-Up for Light Flux Measure m ent ...... 89
21 Luminous Flux Produced by the Slide Projector with and without the Slide at Different Variac S ettin g s ...... 91
22 Luminous Flux Produced by the Left Scene Projector with and without the Film at Different Variac Settings ...... 92
23 Left Screen Luminance without Slide and with Slide (Target Image) at Different Variac S e ttin g s ...... 93
24 Left Scene Luminance without Film and with Film (Back ground Scene) at Different Variac Settings . 94
25 Target Average Luminance and Target Visibility at Differ ent Variac Settings...... 96
26 Display D im ensions ...... 99
27 Three Displays Producing Different Area for Search. . . . 102
28 Schematic Diagram of the Subject's Console, Front Screen and the P rojector ...... 103
x List of Figures Continued
29 Schematic Diagram Showing the Interface between Lab oratory Elements and the Com puter ...... 106
30 Schematic Diagram Showing the Visual Field Determina tion ...... 107
31 Schematic Diagram Showing the Overview of All the Experimental Phases of this Research ...... I l l
32 Schematic Diagram Showing the Computer Control for Data Collection in Reaction Time E xperim ent...... 116
33 Schematic Diagram Showing the Computer Control and • Data Collection in the Foveal Task Experim ents ...... 119
34 Mean and Standard Deviation of Response Times as a Function of Stimulus on Time for Four and Two Digit Displays for Subject SOI ...... 123
35 Cumulative Distribution of the Threshold Visibility Data for Ten Subjects ...... 128
36 Cumulative Distribution of PDA Data for Two Days {Stage I ) ...... 132
37 Cumulative Distribution of PDA Data for Subject SOI (Stage II) under No Load and Load C onditions...... 134
38 Experimental Design for Stage HI Experim ents...... 135
39 Mean PDA as a Function of Foveal Load at Three Dif ferent Target Visibility L e v e ls ...... 137
40 Cumulative Distribution of PDA Data for Three Visi bility Levels (Stage III, Subject S03) ...... 138
41 Cumulative Distribution of PDA at Three Different Foveal Loads (Subject S03, Stage HI) ...... 139
42 Experimental Design for Stage IV Experiments 141
xi List of Figures Continued
43 Mean PDA as a Function of Variac Settings for Four Dif ferent Targets (Slides) Pooled over Two Loads (Subject S 0 3 ) ...... 143
44 Experimental Design for Visual Field Experiment 1 .... 144
45 Mean PDA at Four Levels of Contrast as a Function of Three Levels of Foveal Load and Probability of Error and Response Overlap in Foveal Task as a Function of Foveal Load for Subjects SOI & S02 C om bined ...... 146
46 Latin Square Design for Testing Three Subjects During Three Sessions with Three Foveal L o ad s ...... 150
47 Latin Square Arrangement for Ordering the Observations within a Period under Three Visibility L e v e ls ...... 150
48 Mean PDA at Three Levels of Contrast as a Function of Three Levels of Foveal Load and Probability of Error and Response Overlap in the Foveal Task as a Function of Foveal L oad ...... 153
49 Mean PDA at Three Levels of Contrast as a Function of Three Levels of Foveal Load and Probability of Error and Response Overlap in the Foveal Task as a Function of Foveal Load ...... 154
50 Peripheral Detection Angle vs Target Angular Speed (Experiment 2 ) ...... 156
51 Plot of PDA vs Probability of Occurrence of Errors in Foveal Task ...... 167
52 Cumulative Distribution of PDA for Three Different Foveal Loads (Experiments 1 & 2 ) ...... 177
53 Cumulative Distribution of PDA by Target Visibility Levels (Experiments 1 and 2 ) ...... 179
54 Cumulative Distribution of PDA by Age Groups for the Low Visibility Conditions (Data from, Experiments 1 & 2) . . . 179
55 Plot of PDA vs Target Speed for the 8 Subjects in Ex perim ents 1 and 2 ...... 151 x i i CHAPTER 1
THE ROLE OF PERIPHERAL VISION IN HUMAN VISUAL PERFORMANCE
Introduction
In any man-machine system, the performance of the human element can be characterized by the following three functions:
information acquisition function, • decision making function, and response function.
In many visual tasks such as driving, airplane landing, inspection of complex electronic assemblies, aerial surveillance, etc. information acquisition func tion plays a critical role in affecting the overall performance of the human operator and the entire system. In such tasks, it is imperative for the system designer to consider the operator's visual perception capabilities and limita tions to produce an efficient interfacing of the human operator with other system components. Current literature contains a sizable amount of data on human visual capabilities and limitations pertaining to foveal vision compared to pe ripheral vision. The objective of this research in a broad sense is to investi gate some specific aspects of human peripheral vision and relate their useful ness in an application area such as automobile driving.
Visual Perception:
Visual perception is generally referred to as the human visual function of information acquisition and processing. As the above statement implies, there are two basic stages in visual perception:
The first stage is the acquisition of information through visual sensation and the second stage is the processing and interpretation of the information input in the central nervous system.
Visual sensation is produced in the Retina of the human eye through the stimulation of the visual receptors by the light energy from the objects in the visual scene. The visual receptors will not be stimulated if the light energy falling on the retina is below some threshold level (absolute threshold of vision).
Visual information processing is carried by the central nervous system (CNS) using the neural signals transmitted by the visual receptors through nerve fibers and the information store in the long and short term memory. Based on other factors such as attention, motivation, experience and task- demands, the processed information may reach the short or long term mem ories and also imitate appropriate response mechanisms.
A model of visual perception useful for this study is described in Figure 1, Detailed discussions on visual perception models, stages, etc. can be found In Leibowitz (1965), Graham (1965), Gibson (1966), Comsweet (1970), Scientific American (1972), Murch (1973), Ditchbum (1973), Haber and Hershenson (1973) „and Davidoff (1975).
. The following points can be noted with reference to the above visual per ception model:
1 . The visual scene surrounding any human operator contains a large number of potential visual stimuli.
2. Stimuli from only a portion of this visual scene (visual field) have the chance of being perceived because of the direction of line of sight, head axis and body position.
3. Radiant energy from the stimuli in the visual field is trans mitted into the eye to form a retinal image of the stimuli. This transmission is accomplished by the eye optics.
4. The retinal image stimulates the visual receptors which transmit neural signals to the visual cortex. The success of this stage depends mainly on the properties of the visual receptors of the retina and the nerve fibers which transmit the neural signals.
5. The information transmitted to the visual cortex is processed and interpreted in the light of the stored information (ex perience, task objectives, etc.) in the long term memory. The resulting information may either be transferred to long or short term memory.
6 . The interpretation process m ayor may not be followed by the initiation of some response. ViSUAL PERCEPTUAL SYSTEM . RL EVU SYSTEM NERVOUS TRAL RTTO A TE CEN THE AT PRETATION CORTEX VISUAL RESPONSE T R A N S M IS S IO N INTO INTO N IO S IS M S N A R T RO SN N INTER AND SSING E OC PR THE TO OF N SIGNALS IO S IS M NEURAL S N A R T AND FORMATION AND EYE THE IUL SCENE VISUAL F EIA IMAGE RETINAL OF BODY HEAD, EYE, RETINAL STIMULATION STIMULATION RETINAL INITIATION TI N IO IT S O P MODEL OF VISUAL PERCEPTION PERCEPTION VISUAL OF MODEL FIGURE FIGURE 1 'LOSS INFORMATION INFORMATION INFORMATION LOSS LOSS INFORMATION LOSS INFORMATION LOSS 3 7. At every stage of the visual perception process, there * is the likelihood of information loss,,
8 . One of the response initiations may be to change the eye or head position to attend to different parts of the visual scene.
9. The eye optics and the retina plus the central processes can be called the visual perceptual system.
Foveal and Peripheral Vision;
The primary distinction between foveal and peripheral vision is based on the angle between the line of sight and the target (stimulus) location in the •visual field. If the target lies in the line of sight, its retinal image is received by the receptors at the center of the retina (fovea). As the angle between the target and the line of sight increases, the retinal image of the target moves into the periphery of the retina causing the receptors in the peripheral retina to be stimulated. Therefore, it should be noted that all the stages in the visual perception model remain the same independent of the type of vision (foveal or peripheral) that is used to acquire information and it will be seen later that only the efficiency of the perceptual process changes. Hence, all the factors that are considered in the systematic study of foveal vision are considered in the investigation of visual perception through peripheral vision. Figure 2 repre sents schematically different types of factors associated with peripheral vision. Recent literature by Kochhar (1974) and by National Research Council - Committee on Vision (1975) have systematically identified all the relevant fac tors and the associated literature with regard to human peripheral vision. (See Table 1 .)
Visual Response Criteria;
in the study of visual perception many different criteria are used for the measurement of human visual capabilities. The most common and widely used is the 'Visual Acuity' as measured by using a Snellen Chart. But the term visual acuity represents a number of different visual response criteria. A use ful discussion on this subject can be found in Rubin (1972, 1969), Riggs (1965) and Moses (1970). Perceptual capabilities can be called as "psychophysical" relationships i. e. relation between a stimulus and a response. Visual stimuli presented to an operator can be classified as either intensities (Luminance, L, in foot- lamberts or millilamberts) or extensities (length or visual angles). Then the response can be measured by the strength of the threshold stimulus i. e, the intensity or the extent of the stimulus that is just strong enough to eli cit a response. The reciprocal of the threshold intensity is called sensitivity 5
ENVIRONMENTAL FACTORS
STIMULUS FACTORS
EYE & HEAD AXES POSITION
AGE & SEX EYE O P T IC S & PHYSIOLOGY OF THE RETINA DRUGS & TOXIC SUBSTANCES
TASK VISUAL PSYCHOLOGICAL FACTORS PERCEPTION FACTORS
RESPONSE INITIATION
FACTORS AFFECTING VISUAL PERCEPTION THROUGH PERIPHERAL VISION
FIGURE 2 Table 1 Factors Affecting Visual Perception through Peripheral Vision ’ - ...... —...... Target Operator Physiology Environment (Stimulus) Variables of Eye Task Psychological
Background Nature Angle between line Pupil Size Number of Arousal Display Size of sight and head axis Tasks Attended Luminance Luminance Accommodation Motivation Objects Facial Features Relative Importance Objects Density Size Retinal Factors of each Task Learning Ability Stations xy/Movlng Age Shape Visual Efficiency Response Requirements Experience Atmospheric Sex Type Illumination Color Adaptation Speed Training Heat/Cold Physiological Effects Accuracy Rain Single/Multiple due to Drugs, Toxlo iUUilwwVUjfMnnotnnv : F°8 Substances Similarity of Targets to Dust Location In the Background Objects Reward/Punishment Toxic Substance Visual Field Fatigue Hypoxia Attentlonal Demands Stationary or Motion Restrictions to Operator Field of View Type of Motion Vibration Oscillatory Acceleration Rotational I Sweep Meridian
Contrast
Flicker i Exposure Duration
a> and the reciprocal of the threshold extent is called acuity. The reciprocal presentation of the above measures provides an arbitrary standardization of response data. But Low (1951), Weymouth (1958) and Bloomfield (1972) have pointed out that representing acuity data as the reciprocal of the threshold ex tent of a stimulus leads to erroneous conclusions regarding peripheral acuity and that acuity measures when represented in terms of visual angles provide a correct understanding of the bv.man capabilities.
The response criteria also depends on the type of response the operator is required to make. For example, Rubin (1972) considers the following kinds of visual acuities:
"Visibility Acuity: Threshold size at which the target is visible (seen).
Resolution Acuity: Threshold size at which a specific de tail within a target can be detected.
Shape Acuity: Threshold size at which the shape of a target can be detected (whether it is a circle, a triangle, a square, etc.)
Orientation Acuity: Threshold tilt at which orientation (vertical, horizontal, etc.) of a tar get can be detected.
Vernier Acuity: Threshold displacement between two lines (arranged in the vernier form) to detect the nonalignment between lines.
Motion Acuity: Threshold angle of movement or threshold duration of motion or threshold angular velocity at which the motion of the target can be detected.
Stereoaeuity: Threshold disparity from the horopter beyond which stereopsis is effective (depth perception).
The third dimension to the response criteria is the extent of the visual field at which a given response can be made. This study is primarily interested in assessing the human capabilities in this dimension. The extent of the visual field is represented as the angle from the line of sight. The criteria may be stated as the threshold extent below which a given response under given stimu lus conditions can be elicited. In summary, one should consider the three basic dimensions (see Figure 3 ), i. e. stimulus characteristics^response type and the extent of the visual field in specifying the response criteria in determining human visual capabilities. In this study the primary consideration is given to the visual field dimension. This criteria provides a basis for presenting the rest of the material in this report.
Visual Field as Affected by the Functioning of the Visual System;
The efficient functioning of the visual system depends on three factors.
1 . the characteristics of the different elements (optical media, retina) in the structure of the eye,
2 . the image characteristics of the stimulus, and
3 . the state of the observer.
The most basic elements involved in the structure of the eye are shown in Figure 4. For a complete description, one should refer to Moses (1970) and Pirenne (1967).
Optical Media of the Eye;
Image Location:
The location, intensity and quality of the retinal image are affected by the properties of the image forming mechanism of the eye (Weale, 1956). It consists of the cornea, pupil and lens. Whether or not the rays from a stimu lus reach the retina depends on the angular displacement of the stimulus from the optic axis. Accordingly, an angle of 104° from the optic axis (see Figure 5) has been reported (Pirenne, 1967) to be the limit of the visual field. One should refer to Moses (1970) or LeGrand (1957) for knowing the procedure of calculating the limit of the visual field.
Image Intensity:
The intensity of the retinal image is dependent on the amount of light passing through the pupillary area. It has been noted by Jay (1961), LeGrand (1967), and NRC Committee on Vision (1975) that the effective pupillary area reduces with the increase in the angle of incidence of light from the stimulus. Representative values of the pupillary area as a function of target eccentricities are given in Table 2 (LeGrand, 1967). It can be seen from Figure 6 that the intensity of the retinal image is reduced drastically beyond 70° of angle of inci dence (eccentricity) (Weale, 1956). Normally the pupil diameter may range 9
RESPONSE DIMENSION
VISIBIL
RESOLUTION
SHAPE
ORIENTATION
VERNIER
MOTION
DEPTH
INTENSITY EX TENSITY (SIZE!
120
VISUAL FIELD DIMENSION
THREE B A S IC DIM ENS IO N S OF VISUAL RESPONSE CRITERIA
FIGURE 3 10
Iris
Cornea. Fovea
Lens I ,,. . . ___ _ 5 Visual axis ------
O ptic nerve Retina
Figure 4 —Simplified Diagram of a Cross Section of the Human Eye
i
|
Figure 5 —The Limit of Peripheral Vision (after Pirenne, 1967) Figure Figure 6 —Variation of T ransm ission through Optical Media as as Media Optical through ission ransm T of —Variation ucin fTagt cnrct (fe el, 1956) Weale, (after ccentricity E arget T of Function Percent Intensity Transmitted .9 .1 0.132 0.114 0.094 5 S - deg e .d * 100 85 75 50 25 0 »OCh - o s ea re A of the Entrance Pupil in Relative Values for Several for Values Relative in Pupil Entrance the of Angle of Incidence in D egrees egrees D in Incidence of Angle O I 0.030 0.25 0.44 0.76 0.94 1 1 1 1 ■ Diameters d and Eccentricities n Eccentricities and d ■ Diameters Tagt Eccentricity) arget (T 8 atrLGad 1967) LeGrand, (after O O O Air-cornea Air-cornea Aqueous-Iens t r 1 30 Table 2 Table .4 .5 0.064 0.053 0.042 .404 0.44 0.44 0.94 0.27 0.44 0.76 0.94 2 4 6 d, mm — I— — 60 Os .90.30 0.76 0.29 0.76 X ' \ o \ * O w v \\ * Q U V o i i i 0.15 0.94
12 from 7.5 mm at low illumination to 2.0 mm at high illumination (Ditchbum, 1973). An optimal value for pupil size is some middle value because a pupil of small diameter minimizes the effects of optical aberrations and a pupil of large diameter reduces the blur due to diffraction of light.
Image Quality;
One of the factors affecting the quality of retinal image is astigmatism, i. e., refractive errors due to changes in the angle of incidence at the corneal and lens surfaces. Table 3 represents the astigmatism of the peripheral image in diopters (LeGrand, 1967). A diopter is a unit used to refer to the refractive power of a lens. A high value indicates poor quality of the image.
Table 3
(after LeGrand, 1967) Astigmatism of the Peripheral Image (in diopters)
SA CITTA L SURFACE TANGENTIAL SURFACE E ccen tric ity Nasal side Temporal side Nasal side T em poral side n, d eg C alc. E xptl. C alc. E xptl. Calc. Exptl. Calc. Exptl.
0 . 0 0 0 0 0 0 0 0 10 0.28 - 0 .0 5 0 0.04 -0.47 -0.44 0 - 0 .1 5 20 0.85-0.02 0.28 0.43 -1.40 -1.33 -0.47 -0.56 30 1.71 0.27 0.85 0.50 —2.81 - 2 .1 9 -1.40 —1.52 4 0 2.84 1.02 1.71 0.75 -4.68 -3.48 -2.81 -2.60 50 4.27 2.39 2.84 1.43 -7.02 -4.35 -4.68 -3.39
Note: Sagittal Surface: the focal surface behind the retina. Tangential Surface: the focal surface in front of the retina.
The total astigmatism on a given side at a given eccentricity is given by the sum of the numerical values for sagittal and tangential surfaces.
Example: Total astigmatism on the temporal side at 50° will be 1.43 + 3.39 = 4.82 based on experimental values. Retina of the Eye: 13
The retina, which lines the inside of the eye, carries the light sensitive photoreceptors (visual receptors). The visual receptors in turn are connected to nerve fibers, all of which jointly leave the eye from about 18° from the cen ter of the retina. These nerve fibers are connected to the visual cortex of the brain through a system of neural elements. The area of the retina where the optic nerve leaves the eye does not carry any receptors and therefore; any image falling on this part will not be sensed by the eye. This arrangement pro duces the blind spot in the visual field. It should also be noted that individual nerve fibers are connected to individual receptors at the center (fovea) of the retina. The number receptors that are connected to individual nerve fibers in creases in the peripheral areas of the retina (see Table 4). The visual recep tors are of two kinds: cones and rods which are differently distributed along the retina (see Figure 7). From Figure 7 it can be seen that the density of cones is maximum at the fovea and reduces to a very low value beyond 1 0 ° on either side of the fovea. The density of rods is zero at the fovea and increases gradually to a peak value at about 20°. Beyond this region the rod density de clines, but still remains greater than the cone density. For representative di mensions of different elements of the eye, one should refer to Ditchbum (1973) and Pirenne (1967).
Ditchbum (1973) classifies different regions of the retina in the follow ing manner assuming all regions to be circular and centered on the visual axis:
foveola or central territory — Region of 0.3° diameter
fovea Region of 2° diameter
parafoveal Region between 2° and 10°
peripheral Region beyond 10°
Table 4
Cones in Retinal Areas
Mean no. of No. of Limits of Area on No. of cones per optic the zones the retina, cones optic nerve nerve (>r»deg) mm2 (calc.) fiber fibers 0-5 6.7 200,000 1 200,000 5-10 19.8 300,000 3 100,000 10-20 77.4 700,000 6 115,000 20-30 121 800,000 15 53,000 30-40 157 900,000 30 30,000 40-50 176 1,000,000 45 22,000 50-60 183 1,000,000 70 14,000 60-70 170 900,000 100 9 000 >70 ( 700,000) (100) (7,000) 6,500,000 550,000 Number of rods or cones per mm’ 100.000 120,000 140,000 160.000 180.000 20.000 40,000 60.000 80.000 Figure Figure 70° 70° eprlo rtn Nasal retina on Temporal 0 5° 0 3° 0 1° ° 0 2° 0 4° 0 6° 0 80° 70° 60° 50" 40° 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° 60° 7 40' : Top View of the Left Eye, and the the and Eye, Left the of View Top : Cones 80' Fovea 20 and Cones acro ss the Retina the ss acro Cones and Corresponding D ensities of Rods Rods of ensities D Corresponding ' Perimetric angle (deg) angle Perimetric Visual axi.s Visual ln spot Blind ln pt "MT spot Blind pi nerve Optic Cones Properties of Rods and Cones; 15
The following is a summary of the properties of rods and cones when they are associated with light energy:
1. Sensitivity during Dark Adaptation:
During dark adaptation, the cones regain their sensitivity much faster than the rods. Dark adaptation for cones is almost complete in 5 minutes while rod adaptation takes about 30 minutes (see Figure 8). Also it can be noted that the rods become more sensitive after 7 or 8 minutes than cones.
2. -The Directional Sensitivity
According to this effect, the visual system is most sensitive to light entering the eye near the center of the pupil and progressively less sensitive as the light enters farther from the center of the pupil (see Figure 9).
The directional sensitivity effect applies only when the stimulus is act ing on the cone system during the first 10 minutes of dark adaptation. When the rods become more sensitive than cones then the sensitivity of the visual system will remain the same regardless of the point of entry of light on the pupil. This effect is also called the Stiles-Crawford effect.
3. Spectral Sensitivity:
The spectral sensitivity of the cones and rods is shown in Figure 10o The spectral sensitivity curve (a plot of relative sensitivity and wavelength of light) for cones represents the sensitivity of the eye when the cones are fully dark adapted but are still more sensitive to than the rods. Similarly the curve for rods represents the sensitivity of the eye when the rods are fully dark adapted. The sensitivity of rods and cones at any given wavelength of light depends on the extent of dark adaptation. It should be noted that the sensitivity of rods and cones to higher wavelengths of the visible spectrum (e.g. red) is approximately the same. But the rods are more sensitive than the cones to the light of lower wavelengths.
Also it can be noted that the wavelength for peak sensitivity of cones is shifted to the right (higher) compared to the wavelength for the peak sensitivity of rods. This shift is called the Purkinje Shift. jc JC c a> c tfi cc £ S 0 5 0 5 30 25 20 15 10 S O
Relative threshold intensity _____ 0.2 0.3 0.4 0.6 0.5 0.7 0.8 0.1 0.9 1.0 4 20 2 4 5 4 3 2 1 0 2 3 4 5 u ______Temporal Rods only. Rods Distance between artificial pupil and and pupil artificial between Distance Time in the dark (min) dark the in Time etro aua ppl Immi pupil natural of center i ______1 i ______i ------■ Cones only ■ Cones 1 ------Nasal ■— h hl iul ytm. system visual whole the cones only. The dotted curve is for for is curve dotted The only. cones Fo Cmset 17 Pg 140) Page 1970 sweet, Com (From Curve Adaptation Dark 8: Figure The solid curves a re for the rods and and rods the for re a curves solid The with respect to the center of the the of pupil. center the natural to respect with of the point of entry of the quanta quanta the of entry of point the of Figure 9: D irectional Sensitivity Sensitivity irectional D 9: Figure Fo onwe, 90 . 142) p. 1970 Cornsweet, (From function a s a threshold in Variation of Cones of 16 Wavelength (nm)
350 400 500 600 700 800 8
7
6 Rods - er *5 5 JS I 4 c Cones
i/i c j 8 1
o 28,000 24,000 20,000 16,000 12,000 Wave number (cm"') Figure 10: The Spectral Sensitivity Curves for Rods and Cones
4. Color Sensitivity
Cones are sensitive to different colors whereas the rods are achromatic. No wavelength discrimination is possible in pure rod vision because the rho- dopsin (the photo sensitive substance) contained in rods reacts in the same manner to light of all wavelengths.
The Basis for Visual Function
Visibility and resolution are the most two fundamental functions of the human visual system. "Resolution depends upon the separation of detail and requires the analysis of all the parts of the input. But detection requires only something be perceived and depends upon the pooling and summation of input" (Newman, 1972). The layout of the rods and cones across the retina helps to achieve better resolution at the fovea and better detection in the periphery. Awareness of a dim target against a bright background is a function of the de tection of the just noticeable difference between two levels of illumination at the adjacent receptors of the retina. Detection of a small bright spot on adark background depends on the absolute threshold to light. Spatial Summation: 18
For a target stimulus larger than a point source, but with a diameter less than a given size threshold, Ricco's law states that the product of target luminance and area gives a constant. Ricco's law applies only for areas small enough to be served by a single receptive field. As the stimulus area increases beyond Ricco's area to some intermediate size, then the product of target lum inance and the square root of stimulus area is a constant. This relationship is called Piper's Law. Beyond Piper's area there is no spatial summation. De tection threshold becomes independent of target size and depends only on target brightness. Spatial summation increases with distance from the fovea. Spatial summation can be mathematically expressed as AL * = constant, where A L is thedifference between the target and background luminances, A is the tar get area and k is referred to as the summation coefficient. K = 1 refers to complete summation and k = 0 refers to no summation.
Temporal Summation:
There is a critical duration of target exposure below which the lumi nance and the time are reciprocally related. This relation is referred to as Bloch's law and is given by AL * T = constant. This constant increases with increased background luminances (see Figure 11). Critical duration is about 30 msec, for the highest background luminance and is about 100 msec, for the lowest background luminance.
Background Luminance
Both spatial and temporal summation are affected by changes in the background luminance. Hence target contrast threshold is affected. Figure 12 shows the contrast sensitivity for a target (10 minutes of arc in diameter) as a function of retinal locus for five background luminances. Contrast sensitivity of the retina ijs said to be high if for a given background luminance the thresh old contrast is low. From the figure it can be seen that the contrast sensitivity of the fovea is highest when the background luminance is at the higher photopic levels. It is lowest at the fovea when the background luminance is at the sco- topic levels. At mesopic background luminances, the contrast sensitivity is uniform over much of the central field (20° nasal to 40° temporal). Towards the periphery the contrast sensitivity decreases. Also, at any given part of the periphery, the contrast sensitivity decreases with the decrease in back ground luminances. The rate of decrease is approximately uniform across the visual field.
Perception of Motion
Movement perception is affected by changes in retinal positions of visual excitations, especially at the leading and trailing edges of the stimulus s/t
<* . 0 THRESHOLD CONTRAST {AL/U 1000 100 0.1 10 I I I lL1 I I I I I I I L I IJ I I I I L I I I I I I I I I I I I I I1 I - l_L I I I I I I I 5 a =a dm. oetitteodnt ersns sensitivity. represents ordinate the tliat Note luminances, backgroundcd/m2. different for locus retinal of function a as threshold Contrast NOTE: 3.183 cd/m ^ = 1.0 mL, - 0.9293 ft, L. ft, 0.9293 - mL, 1.0 = ^ cd/m 3.183 NOTE: cd/m2; 2 (From Committee on Vision, National Research Council) Research National Vision, on Committee (From 02 c/2 c= .2 c/2 d- 002 c/2 ea ,00027 0 a e cd/m2} 0.0027 - d cd/m2; 0.027 = c cd/m2j 0.27 = b DISTANCE PROM THE FOVEA IN DEGREES IN FOVEA THE PROM DISTANCE 0 2 40 20 0 20 iue 12 Figure temporal visual field visual temporal 10' te tt tt te 10' *001 *001 0* meridian 0* *B.P. By* t ljh R 60 80
o to 21 image. If the stimulus is a point, the component effects centered at successive loci as the stimulus moves in the visual field may be thought of as constituting a contrast wave. Stimulation of retinal receptors becomes more complex when the stimulus is not a point source. Generally velocity threshold for perception of movemont of a line decreases in the following situations:
1. Increase in the number of reference objects
2. Increase of stimulus line length
3. Increase in stimulus intensity
Graham (1965, 1971) reports that the conditions of luminance (L) and duration (t) required for motion perception follow the Bunsen-Roscoe Law, i. e. Lt = constant, for t up to 0.3 seconds. At longer durations, luminance alone deter mines the velocity threshold.
Literature Review on Peripheral Vision
In an attempt to identify the variables that need to be researched in the area of peripheral vision, available literature was reviewed and is summarized in the following sections. Table 1 provides detailed identification of all the variables affecting performance in the visual periphery. In this section litera ture on some of the important variables are reviewed and discussed.
Performance Measures
Different performance measures were used in different experiments to suit the instrumentation requirements, hypotheses to be tested, etc.
Response Time
This measure was mainly used for responding to stationary peripheral targets (with given angular location). Response time was given by the time duration between the on-set of the peripheral stimulus and the time of response.
Accuracy of Response
This measure was used when the possibility for making errors in re-- sponse such as misses and false alarms existed.
Threshold Target Luminance or Target Contrast
This measure determined the subject's threshold for target luminance or contrast at which he can first detect the peripheral target. The threshold 22 type of performance measures can in fact be applied to target variables other than luminance and contrast such as size, velocity, etc,
Foveal Task Demands
Peripheral response angle is always measured from the line of sight or foveal fixation point. Hence, it becomes of primary importance to consider de mands imposed at the fovea of the retina when studying the performance of the periphery of the retina. The degree of task demand at the fovea may be direct ly responsible to subjects1 level of attention to objects at the fovea. Moray (1970) identifies the following broad categories as the major subdivisions of the concept of attention:
1. Mental Concentration: The person concentrates on tasks such as mental arithmetic and tries to exclude all incoming stimuli which might interfere with the performance of the specified task.
2. Vigilance: This is a task in which the observer is paying attention in the hope of detecting some event whenever it happens.
3. Selective Attention: It is the task of selecting only one out of several messages to accept and respond to.
4. Search: It is the task of hunting from a set of signals for a subset of signals or a single signal.
5. Set: This is the task of preparation to respond in a cer tain way either cognitively or by motor responses.
Depending on the specific foveal task involved, each situation can be classified as having some type of attentional task at the fovea. Table 5 summarizes some of the important studies which investigated the effect of different levels of foveal task or demands on performance in the visual periphery.
Tracking task involving eye hand coordination was used to compete for attention with peripheral stimuli by Gasson and Peters, Kephart and Chandler, Robinson et. al., and Goolkasian and Bunt. The results show that there is a general tendency for the visual field to shrink. Also in situations where the eyes have to move to the peripheral stimuli, the reaction time will increase in the persence of the central task.
Signal monitoring task such as counting the number of blinks or reillum- inating the extinguished light by pressing a button was used as a foveal task by Table 5
Summary of Literature on the Effect of Foveal Task on Performance in the Visual Periphery
Reference Foveal Task Peripheral Task Result
Abemethy and Leibowitz A centrally fixated light was Luminance of periph With increases inter (1971) extinguished at random inter eral signals were in ruption rate, eyes fix Leibowitz and Apple vals. The subjects were re creased until subject ate for a higher percent (1969) quired to press a button and reported detecting of time at the central Leibowitz (1971) reilluminate the light so as to them. Location of task. Reaction time to keep it on always* Levels of signals: 90°, 80° , peripheral signals were difficulty: Approximately 28 65°, 50°, 35°, and not affected by central or 59 p resses per minute were 20° on either side. task difficulty. During required to keep the light il initial sessions, thresh luminated. old luminance for tar gets were higher with- higher interruption rate in the central task. With practice the threshold luminance at both levels of central task were the same.
Gas son and Peters Steer a pointer along a spiral Identify the side Significant reduction of (1965) path on a rotating cylinder with (left or right) in the visual field under the eyes fixated at the center. which signals oc- foveal task. curred. to CO Table 5 (cant’d)
Reference Foveal Task Peripheral Task Result
Goolkasian and Bunt Tracking task: subjects were to Prediction task: A In the dual task situation (1973) compensate a horizontal move peripheral light moved prediction of collision ment of a circular light by means towards the central suffered when the periph of a joystick control. Under per light. The subjects eral lights appeared fect compensation the light should had to predict whether from 45° to 30°. move vertically up in a straight or not the two lights line. would collide at the center. The periph eral light was pre sented from 1) 33° to 15° or 2) 45° to 30°.
Huntley (1973) A central light was interrupted Detect and locate Reaction times to pe by one, two or three closely the side of a light ripheral signals were spaced interruptions. Subjects signal presented at increased with an in were required to count and re 4, 24, 44, 64, and crease in foveal task port the number of each type of • 84degrees0 difficulty. signals at the end of the trial.
Kephart and Chandler Steer a pointer within the limits Identify the side Significant reduction in (1956) of a moving simulated road. (left or right) in the visual field in the which a white but presence of the track ton appeared. ing task,
to Table 5 (cont’d)
Reference Foveal Task Peripheral Task Result
Moskowitz et, al. Count and report the number Detect the periph Percent misses in the (1972) (1974) of blinks of a centrally fixated eral light signals peripheral light detec light during each trial. Blink placed from 12° to tions increased with an rate: 0, 0.4, and 0.8 per 102° at 6° intervals. increase in the blink second. rate of the central light. Marihuana and central load did not interact. The reaction time to peripheral lights when detected was not af fected by central load. Significant alcohol and central load interaction was observed.
Robinson et. al. Central manual control task Upon a command In the presence of the (1975) simulated direction and speed signal subject in central task head moved control. terrupted central first while eyes made task and moved his compensatory move visual fixation to a ment and stayed on the peripheral digital central task. Mean eye display and re reaction time with and sponded to the dis without central task play. (Peripheral were 634 and 175 msec. target location - 100° to 130°). to Table 5 (cont’d)
* Reference Foveal Task Peripheral Task Result
Robinson et. al. A pursuit control task: sub Detect and discrim In the presence of the (1975) jects were required to keep a inate a peripheral central task head move dot within a given circle on a lighted digit located ment preceded eye move TV display. at 30° and 90° from ment and the eye move the center. ments showed compen- satoiy movement which were not observed in the case without the central task.
Weltman and Egstrom Two central tasks were used: The subjects were Mean response time to (1966) 1) addition task and 2) dial to detect and extin peripheral signals sig monitoring task. In the addi guish a light placed nificantly increased in tion task a row of digits to be at 60° on the left pe the presence of the added and circle the digit which riphery. Signals central task when per brought the sum to a reference presented at random formed under water. value. The dial monitoring intervals of 25 to 65 task was a vigilance type of seconds. Signal rate task in which one has to identi was 75 per hour. fy a 75° deflection among the many 45° deflections of the dial. The dial deflected once every second.
to « Table 5 (cont'd)
Reference Foveal Task Peripheral Task Result
Zahn and Haines Search for a given target in a Detect the on-set of With the increase in (1971) 5 x 5 matrix of letters and fix any one of 7 periph search task luminance ate for a second on the target,, eral lights placed at time to detect periph Central display luminance was 90°, 60°, 30° on eral targets increased. kept at 4 levels (8.5, 55, 792, either side. 6800 ft. L) 28 pressing a button was used as a foveal task by Moskowitz et. al., Huntley, Aber- nethy and Leibowitz, and Leibowitz and Apple. Conflicting results about re sponse times to peripheral signals have been reported.
Vigilance type of foveal task was used by Weltman and Egstrom and the mean response time to peripheral signals were found to increase.
A search task was used by Zahn and Haines and it was observed that in creased luminance on the search task increased the time to detect peripheral targ ets.
Sanders (1963, 1970) defines the following three fields:
1. Stationary Field — mere peripheral viewing leads to efficient performance.
2. Eye Field — supplementary use of eye movements is required.
3. Head Field — where head movements also are required.
Sanders investigated the transition region from one field to another under dif- ferent types of perceptual loads. Effect of central task demands can be studied by observing the changes in the transition region from the stationary field to the eye field. From Sander’s experiments, it can be seen that the above tran sition extends from 31° to 52° under low perceptual load conditions and from 19° to 34° under high perceptual load conditions.
The general conclusion from all the above investigations is that any contraction of the visual field during increased foveal attention is a psychologi cal effect involving cortical control over conscious perception and not due to retinal activity. Hence, it is recommended that the type of foveal task used be selected such that it demands suitable sensory behavior corresponding to the application area. For example, in driving, visual search for information may be the primary task followed by tracking (lateral and longitudinal control of the vehicle).
One important aspect that was missed in all the previous experiments using foveal and peripheral tasks is the concept of subjects' spare capacity for performance. Since each subject may vary in his spare capacity to carry an additional task, a given level of foveal task demands may not produce the same level of attention at the foveal task. To avoid this problem: 29 1. The homogeneity of subjects may be maintained with respect to spare capacity. o r 2. The foveal task load may be adjusted for each sub ject so as to load each subject according to some loading criteria.
Following the above conditions for the foveal task may ensure reliable results on the peripheral task performance.
Peripheral Attention
The previous section discussed the effect of foveal attention on the per formance in the visual periphery. In this section few aspects related to periph eral attention above will be discussed.
■ Babington Smith (1961) and Thompson (1961) have reported that if one tries to concentrate on selected objects in the periphery with the eyes fixated at the center, the object selected for attention becomes hazy and may even dis appear. It is suggested that some form of central interference is responsible for this phenomenon.
Mackworth (1965) observed that the performance in a letter recognition task presented in the periphery was impaired when the letter was presented with extra letters. It was concluded that extraneous visual stimuli destroyed peripheral recognition by producing complex stimulus patterns in the periphery.
Reeves and Bergum (1972) hypothesized that "under aroused states the individual will attend to need related cues regardless of where, or in what com bination they occur in the visual field". Based on their experiments, the au thors concluded that nonspecific arousal complicates the discrimination process leading to declined peripheral detection efficiency. But under specific arousal conditions involving relevant peripheral cues detection response is improved.
Under conditions of arousal Cornsweet (1969) also observed that sub jects were attentive to relevant information occurring in the visual periphery.
Schioldborg (1971) investigated the effect of focusing and dispersion of attention on the visual identification time for alphabets in the central vision and periphery under the following conditions:
a) with and without foreknowledge of target position
b) either above or with other digits 30 The following were the results:
1. In both central and peripheral vision, response times were longer when the target was presented with other objects (increased total stimulus un certainty).
2. With the foreknowledge of position (reduced un certainty about target location) the response time was decreased only when the letters were pre sented with digits.
A study by Grindley and Townsend (1968) investigated whether voluntary attention to a particular part of the peripheral visual field had any effect on the accuracy of the subject's perception. The results showed that
• 1. When the peripheral target alone was presented, reduced uncertainty about target location did not improve the peripheral response accuracy.
2. When the target was presented among ojtl^er ob jects, foreknowledge of the target position sig nificantly improved the accuracy of peripheral responses.
The overall conclusions regarding peripheral attention can be summarized as follows:
1. Detection of peripheral targets under foveal stress or any kind of arousal greatly depends on the rele vancy of peripheral stimuli.
2. When the target is to be viewed among many back ground objects, the subject's attention is dispersed. Therefore, the momentary attention level for the tar get is lowered. Hence, focusing and dispersion of attention may be considered antagonistic processes.
3. During attention, localization and identification take place at the same time. Identification normally occurs in the central vision and the corresponding momentary attention level is highest. Therefore, a dispersion of attention increases the momentary attention level for peripheral objects only to the ex tent of locating the objects and not for identification. 31 The general conclusion can be captured in the following table.
Attention Attention Focused D ispersed
(Increases momentary (Increases momentary attentional level in the attentional level in the central field central field
Identification in the good poor central vision
Localization in the poor good periphery
Peripheral Motion Detection
Among many target variables motion has been investigated in the periph eral vision context only by very few investigators. Among them, the work of Johnson and Leibowitz (1974), Leibowitz and Johnson (1972), Low (1947), McColgin (1960), and Rogers (1972) may be considered as important.
Low (1947) used moving Landolt rings at angular speed 15 degrees per second and exposed the target for one second duration at angular locations of 75°, 60°, and 45°. The major results were as follows:
1. Peripheral motion acuity does not differ markedly between right and left eyes.
2. The effect of direction of movement of targets (in ward, outward) on the peripheral motion acuity is very little.
3. Motion acuity declined sharply from 30° to 60°.
4. The peripheral motion acuity for a 15° exposure (from 45° to 30°) was not very much different from that for a 3° exposure (from 33° to 30°).
5. The following table shows the comparison of motion acuity and stationary acuity results. 32
Mean Acuity (mm of Landolt Rign Gap Recognized)
Motion Acuity Stationary Acuity (Low, 1943)
*4 & 30° 2.65 1.6
JoS 0 c o to o w ell over < cu 8 .0 5 .0
From the table it can be seen that the motion acuity is weaker than the stationary acuity at given angular loca tions of the target (30°, 60°)o
6. Motion acuity scores did not show appreciable improve ment through practice.
McColgin (I960) conducted a study to provide some basic data on thresh olds for movement in peripheral vision. The subjects were required to respond to the movements of a standard altimeter hand in an aircraft type instrument located in the periphery. The altimeter hand either rotated or reciprocated. Movement thresholds were determined by method of limits. The dependent variables were revolutions per minute for rotary motion and strokes per minute for linear motion. The important results are given below (Figure 13).
14
270 90 2 70 W 9 0
ISO Perimetric. ___ _ chart , showing i r ” the absolute threshold . Perimetric chart showing the absolute threshold iso- jsograms (rpm) of rotary motion. grams, m strokes/min, of linear motion. Vertical motion is repre sented by solid lines and horizontal motion by dashed lines. Figure 13: Data on Movement Thresholds (After McColgin, 1960) 33 1. The general result from the figures is that the iso grams form concentric elliptical patterns with the horizontal axis approximately twice as long as the vertical axis.
2. An individual's sensitivity to movement steadily de creases as a linear function from the fovea to the periphery.
3. The direction of rotation (clockwise or counterclock wise) did not affect the subjects' perception of move ment in the peripheral vision.
4. An individual's ability to perceive vertical motion is slightly better than his ability to perceive horizontal motion in the area adjacent to the horizontal meridian.
Johnson and Leibowitz (1974) investigated the effect of practice, cor recting peripheral refractive error, feedback on performance and longevity of practice and feedback upon movement thresholds in the periphery. The sub jects reported whether a (white, square) stimulus in the periphery moved to the left or to the right or remained stationary. The major results are given below.
1) Practice has little effect on movement threshold at eccentricities less than 20°, whereas movement threshold greatly improves beyond 20° eccentricity.
2) With the correction for refractive error, move ment threshold significantly improved (lowered) at every stimulus location except for the fovea. Also the correction produced very low individual differences.
3) Feedback improved movement threshold in the pe riphery only when the refractive error was not cor rected. But the results also indicate that feedback with correction for refractive error improves move ment threshold when compared to feedback alone.
4) Movement thresholds after 3 months were com parable with that with practice. 34 The author concludes that:
1) Feedback serves to improve the interpretation of a degraded image.
2) Correction of refractive error improves the retinal image quality.
3) Improvement in movement threshold with practice is not a short term effect.
Peripheral Contrast Thresholds
Target contrast is another target variable which has not been investi gated very well in the experimental context of peripheral detection. Rogers (1972) investigated the difference in contrast threshold for moving images be tween foveal and peripheral stimuli. Angular speed of the target was 24°/ second and the stimulus was viewed at 0, 20, 40 and 55 degrees. The results from the study were as follows:
1. Image brightness threshold was lower for dynamic stimulus when compared with stationary stimulus (for foveal and peripheral viewing).
2. Interaction between stimulus velocity and eccentricity was observed.
a) With the stationary stimulus, the threshold image brightness increased with an increase in eccen tricity .
b) With dynamic stimulus, the threshold image brightness was not affected by stimulus eccen tricity (55°).
3. The direction of the stimulus movement had no effect on the threshold image brightness.
In conclusion, the perceptual sensitivity of the retina is greater for an image in motion than for a stationary image.
Eye-Head Movements
The detection of objects in the visual field is highly dependent on the eye-head position with reference to the objects. Peripheral vision is considered 35 to be responsible for planning the eye movements. Hence, the literature on eye-head dynamics is reviewed briefly here.
Bartz (1962) investigated the eye movement latency (eye reaction time) to peripheral signals as a function of angular position of signals. The results were as follows. The total response time to signals increased with an in crease in angular location. Visual reaction time ranged from 0.2 to 0.24 seconds.
Bartz (1966) also reported that eyes move faster than the head in re sponding to a peripheral signal. Initially the eyes were fixated at a central point. Bartz (1967) reported that more undershoots of eye movements during fixation occurred with an increase in peripheral angle of stimuli. Robinson et. al. (1975) confirmed Bartz's findings on eye-head movements during re sponse to a peripheral target without any central task. But in the presence of a central tracking task, Robinson observed that head movement occurred first and the eyes moved later toward the target. This result shows that the eyes compensate for the slow movement of the head and remain in the central task until the end of eye reaction time. Robinson also reports that the time loss due to eye-head dynamics during visual search can be expressed as (84 + 3.50) msecs., where O is the angle of target location.
Studies on Visual Search
Discussions on eye-head dynamics logically leads to the consideration of vast literature available in the area of visual search and inspection. It is not the intent of this report to review all of them. But a few important recent studies will be reviewed to highlight the need to understand the human periph eral vision capabilities.
The most recent literature in the field of visual inspection and visual search may be considered to be by Bloomfield (1969, 1972, 1974), Williams (1966, 1967, 1970), and Gould (1969, 1974). Summaries of each of the studies are reported briefly here.
Studies by Bloomfield
Starting from 1968, a series of research papers in the field of visual search have been published by Bloomfield. In his research, Bloomfield con sidered the mean search time and the response time as the performance mea sures and related them to the target characteristics such as size, shape, color and texture and the relevancy of the background to the target. He also investigated the effect of viewing distance on peripheral acuity with complex stimuli. Based on the evidence from his research, Bloomfield and Howarth proposed a theory of visual search relating human characteristics, target de tails, target background and search time. Important results and the major conslusions of Bloomfield's research are summarized on the following page. As the difference in size of the target and nontargets decreases, the frequency distributions of times needed to locate the targets increase in their variability and range.
The distributions become increasingly skewed.
The means and the medians of the distributions increase.
When the difference in size between the target and the nontargets is very large, the targets are perceptually prominent. In this case search is not necessary. The cumulative distribution of search times will be very steep.
The assumption that the search times are exponentially distributed is supported only for the cases where search is involved to a higher degree i. e. size difference be tween the target and the nontargets is very much less.
Search time decreased markedly with a decrease in the density of the nontargets in the display.
The results support the square root relation between search time and the number of nontargets.
t = a. N1/2 , a is a constant N is the number of nontargets
The results support the following relationship between the search time (F) and the target-background charac teristics:
(dB " dT)2 where dg and dy are the diameters of the background and target stimuli.
The results support the following relationship between the response time (tf) and the characteristics of the target-background; 37
a ------|d B - d x |
where the response time tj is the time to detect the target at the first glimpse with no element of search,
6. The mean search time for a target in a plain back ground can be given by the relation
(dT - d Q)2 where dQ is the threshold diameter of a target at the fovea,
7. It is suggested that an INDEX OF TARGET DIFFICULTY could be developed by treating any target-background complex in terms of size measures by using the rela tions for F and tf described in 4 and 5 above. ! ! 8. Detection of one of the targets in a multiple target situation is faster when
a) the number of targets in the display is higher.
b ) the targets are grouped adjacent to each other compared to the random dispersion throughout the display,
9. When targets of shapes: square, hexagon, octagon and duodecagon were used in a search experiment, the re sult was that the search time increased with the target shape becoming more like a disc (from square to duodecagon).
10. Rated discriminability of color targets embedded in a yellow background was studied. The results show that the following is the order of colors from easy to most hard discrimination:
Red Blue Green Tan White 38 11. The visual search data for the above targets showed that red and blue targets required very little search. But green, tan and white targets required increased search time.
12. The peripheral acuity data showed the mean distance (in minutes of visual angle) away from the fixation point at which the different color targets were de tected. The results are as given below:
Red Blue 11 degrees approxim ately Green 8 Tan 4 White - 3
13. The results from the search for the embedded color targets support the following relations developed by Bloomfield:
t a 1 and t a 1 D2 O 2
where F is the mean search time
D is the score of the rated discriminability
© is the extent into the periphery that a target can be detected.
14. Visual search studies with texture differences did not support the above relations.
15. In another study involving visual search at two viewing distances (7 and 47 feet) it was found that the relation ship between search time and peripheral acuity was in dependent of viewing distance.
Visual Search Theory (Bloomfield and Howarth):
The following are the final equations for the mean search time for dif ferent types of search: 39
COMPETITION SEARCH PLAIN SEARCH
REGULAR A • • P, SEARCH
RANDOM SEARCH — f t (d T “ d o )
T is the mean search time.
t„ is the average fixation duration.
pg is the probability of detecting a target in a single fixation.
pa is the probability that a target will be detected if it falls in a certain area around the line of sight.
A is the search area.
m is the slope of the graph obtained by plotting the minimal visual angle detected vs. the angular distance away from the fovea at which detection is made.
dp is the diameter or the size of the background objects.
drp is the diameter or the size of the target.
dQ is the threshold diameter of a target on the line of sight.
Studies by Williams:
Williams (1966, 1967, 1970) studied the visual search of human observ ers In fields which were crowded with objects. The basic framework in his studies which used eye movement records was that the basis for fixating objects in the field was some characteristic of the target being searched. The major conclusions from his studies are summarized below: A theoretical model for the search time distribution was developed and was shown to approximate to an exponential distribution.
Search is not random, but it is based on some known characteristic of the target.
When information about more than one characteristic of the target is available, the subjects tended to use only one characteristic as the basis for search.
The following is the decreasing order of preference of variables for use as the basis for search: color, size, and shape.
The concept of similarity of objects to the target was developed based on the proportion of eye fixations for each specified level of the target characteristics.
Five shapes: circle, semicircle, triangle, square and rectangle were studied for similarities among the theme Circles and triangles were fixated more con sistently.
Shape similarity was not affected by the density of objects.
Extrafoveal discrimination of shape was slightly im proved when all the objects were of the same size and contrast.
Square objects of different sizes, with different con trast were used to study the problem of size similarity. The similarity of a given size object to the target de pended only on the ratio of the two sizes. It was con cluded that there is a discrimination gradient for size that is relatively unaffected by other measurable char acteristics of the field.
The mean search time was significantly affected by the type of information on target characteristics that were available to the observers. When color information was provided the mean search time was about 6 to 8 sec onds. With size and without color the time was about 16 seconds. With shape alone the time was about 20 seconds. Without any basis it was about 23 seconds. 41 11. The data on search times was highly related to the proportion of fixations on specified target information.
12. Increase in the object density in the field increases the search time.
13. Efficiency in search is related to the observer's extra foveal discrimination of different target characteristics.
14. A prediction equation for the mean search time was proposed relating object similarities to the target, ob server’s fixation behavior and the probability of locating the target.
It should be noted that the eye movement data was used to determine the number of fixations only. It was not used for temporal analysis.
Studies by Gould
Gould considers eye movement records as useful in theorizing about cognitive processing of visual information. According to Gould, the four char acteristics of eye movements and the factors affecting each of the characteris tics are given in Table 6. Gould also has developed a model of the on-going processes during an individual eye fixation in a visual search task. The model is described schematically in Figure 14.
All the three investigators whose work was briefly reviewed earlier have stressed the importance of human peripheral detection capabilities in the performance of visual search tasks. In the model for search time, one of the inputs is the extent into the periphery a target can be detected under given tar get and background conditions. The literature also suggests the relevancy of an object plays an important role in being detected and fixated during search. The above investigators also stress the need for improved understanding anri sufficient data on human peripheral detection through visual systems. 42
Table 6 Characteristics of Eye Movements Four Characteristics Factors Affecting the Eye of Eye Movements Movement Characteristics
1. Duration of Observer's Experience Eye Fixations Complexity of the Task Degree of Similarity between the Fixated Object and the Target
2. Location of Likelihood of Target Location on Eye Fixation the Display Observer's Knowledge about the Char acteristics of the Search Target Intention or the Purpose of Search Observer's Experience
3. Sequence of Nature of Search Task Eye Fixations Similarity between Background Objects and the Target
4. Useful Visual Noise Visual Field Target-Background Similarity Quality of the Display
« 43
J PR E - This process occurs before the beginning of the 1 A TTENTIONA L first eye fixation. Objects are separated from 1 PROCESSES the background based on the observer's knowl edge about the target characteristics and his in tention or purpose of search. This process can be called as the psychological structuring of the search field or the perception of a selective pat tern in the search field.
FOCUS At the beginning of each eye fixation, the image IMAGE is blurred for a short duration due to over- or under-shoot of the eye movement. Therefore, this adjustment of the eye takes place.
FEATURE Critical features of the fixated object are ab ABSTRACTION stracted and encoded. The time for this process AND depends on the degree of degradation of the dis ENCODING play. Target-object similarity does not affect this process. Information abstracted is much more when compared to the amount used.
FEATURE Encoded features are compared with the mem COMPARISON orized target images. The time for this com parison depends on the target-object similarity. Time increases with increased similarity. With multiple targets the observer may compare simultaneously or sequentially depending on his training in the task. The error rate is less in the sequential comparison than in the simulta neous comparison. The search ends if based on the comparison it is decided that the object is completely similar to the target.
PREPARE Just prior to the termination of an eye fixation, TO there may be a reduction in the visual sensitivity. REFIXATE ----- 1 r NEW ! FIXATION Figure 14; Gould's Model of the Ongoing Pro cesses during an Individual Eye Fixation in a Visual Search Task BLANK PAGE CHAPTER 2
ROLE OF PERIPHERAL VISION IN DRIVING
Introduction
Among many tasks in driving, drivers constantly obtain visual informa tion from the road straight ahead as well as from other directions. This infor mation may be related to
1. The traffic ahead
2. The roadway ahead
3. Sign information
4. Other vehicles and highway targets which are on a collision path with the driver's vehicle.
While the information sources are too many, the driver can fixate his eyes at any given instant in one direction only. The time duration of fixation in any given direction or at any source of information depends on
1. The relevancy of the information to the driving task
2« Information uncertainty
3. Degree of details contained by the source
4. Capacity of the driver's visual system
Not all bits of information available on the highway are required by the driver for his driving. Also the relevancy of a bit of information may change from one maneuver to another. When changing lanes or merging, traffic in the adjacent lanes becomes more relevant than during normal driving. Normally, driver's eyes have the straight ahead position as their home base. Whenever relevant information from other directions has to be.processed, the eyes and/or head move and the driver begins to share time between two or more sources of infor mation. In fact, driving involves a constant time sharing between different in formation sources. Some sources such as highway signs contain detailed infor mation and require direct foveal fixation and resolution of details by the drivers. 45 46 But in situations such as lane changing, the driver needs to know only the pre sence of any other car in the adjacent lanes. In the latter case, a simple de tection of the target will provide the required information to the driver.
The necessity for movement of the eye and/or head arises based on the nature and location of the information to be gathered and the capacity limitations of drivers1 visual system. Drivers' normal field of vision extends to about 90° to 100° on either side. Within this field, only small central region details can be resolved effectively. The resolution capacity decreases toward the periph ery. Hence, a driver may be forced to make head and/or eye movements if the source of information is located beyond his visual field. The above movements can also occur if the driver wants to increase the chances of resolution or de tection of the detail by reducing the eccentricity angle and increasing the clarity of details,, The automobile designers aim to provide as much visual field as possible by their mirror system. In this process they try to develop standards for the location of mirrors. The highway designers try to increase the sight distance at intersections and develop better geometry for merging so as to in crease the probability of detection of the targets. Efforts from both groups of experts will require inputs regarding human capabilities in responding to periph eral targets in different driving environments.
Foveal and Peripheral Tasks in Driving
Generally, foveal task refers to the information processing at the eye position and the peripheral task refers to the processing of information avail able at eccentric positions. During driving, the eye position changes frequently and so are the eccentric angles of the information sources. Changes in the eye position are mainly due to the different types of visual search behavior of dri vers under different driving activities. Table 7 lists some of the common dri ving activities and their associated foveal task and peripheral tasks and targets.
Table 7 describes only three of the many driving activities with the asso ciated foveal and peripheral tasks. In any given driving situation, the factors that affect the drivers' useful visual field can be identified under the following factors:
1. Driver: -Efficiency of Visual System -Normal Visual Field -Age -T raining -F atigue 47 2. Foveal Task Demand: -Objective and Purpose of Driving -Driving Activity -Traffic Conditions -Driving Speed -Familiarity of Route
Peripheral Target Variables:
Nature: -Car, Truck, Pedestrians, Signs, Signals, etc. Motion: -Stationary or Moving Color: -Bright or Dim Size: -Large or Small Location: -Angular Location in the Field of View Track: -From behind, Oncoming, from the Side T arget Visibility: -Target Luminance -Background Luminance
4. General Lighting Conditions: -Day, Night, Dawn, Dusk, and Street Lighting
5. Weather: -R ain, Snow, etc.
6. Road Geometry: -Sight Distance, Alignment, Merge Angle
7. Automobile Design: -View from Inside the Automobile, Mirror Location, Pillars, etc.
Literature dealing with the effects of the above factors on drivers’ pe ripheral response are very limited. Some of the important ones which are directly related to driving have been reviewed and reported in the following sections. 48 Table 7
Foveal Task and Relevant Peripheral Targets for Some Common Driving Activities
Activity Foveal Task Relevant Peripheral Targets & Tasks
Normal Driving 1. Detection of vehicles passing in with eyes Monitoring the adjacent lane from behind Straight Ahead Straight Ahead Traffic 2. Detection of merging vehicles 3» Detection of vehicles in the crossroads 4. Detection of traffic signals, signs, pedestrians
M irro r Gather Information 1. Detect changes in the motion Sampling about Rear and Side of the traffic ahead Traffic 2. Detection of changes in the headway of lead vehicles 3. Detection of passing and merg ing vehicles 4. Detection of traffic signs, sig nals, pedestrians, etc.
Highway Sign Search for Required 1. Detection of traffic ahead and Reading Sign Details at crossroads 2. Detection of lane changing and merging vehicles Visual Field Requirements in Driving 49
In driving, drivers constantly search for relevant information. Periph eral vision is one among many factors that are responsible for initiating eye movements leading to visual search. It can be noted from visual search litera ture (Erickson (1964), Johnston (1965)) that good peripheral vision leads to effi cient eye movements, shorter search time and hence, enhance detection of tar gets. Hence, it is considered (Allen (1970) that good peripheral vision is essen tial to driving which involves frequent visual search. Currently there are no uniform standards for examining the visual field of drivers at the time of driver testing for licensing. Roberts(1971) considers that a normal field less than 70° on either side may produce a failure to recognize driving hazards in the periph ery. A sa rule of thumb, an individual should be able to perceive the movement of a finger or other action 90° or more away from his direct forward line of sight. Allen (1970) considers that a person having visual fields less than 70° on either side requires medical help. Early useful discussions on this subject have been provided by Danielson (1957) and Kite and King (1961). The Traffic Engi neering Handbook (1965) reports that only 12 states in the country required the drivers to have a minimum specified visual field in degrees. This minimum value ranged from 90° to 150° on both sides.
The Highway Safety Research Center at the University of North Carolina (1974, 1975) tested the visual fields of about 52, 000 drivers during driver li censing using Ortho-raters. Two years accident experience of those with lim ited visual fields (140 degrees or less) was compared with those with normal fields of view (greater than 160°). The study indicated that narrower visual fields may not be related to significantly higher accident involvement. The study also reported that visual fields of 120° or less were observed with only less than one percent of North Carolina driver population. A higher proportion of older drivers had limited visual field. It was also found that restricted visual fields may be slightly related to a higher proportion of side collisions. How ever, based on the conclusion in the above investigation, it can be stated that limited visual field of drivers may not lead to accidents. Accident statistics are usually considered as the ultimate criteria for measuring the effectiveness of any safety countermeasure. 'Because of the complexity of factors involved in accident causation, it is difficult to identify the factors that cause accidents. Hence, suitable intermediate criteria are required for evaluating the effect of limited visual fields of drivers.
Automotive designers consider the drivers’ visual field requirements as an important input in developing efficient rear vision systems for the automo bile. Forbes (1970) considers the information on the relative importance of various fields of view and the inherent visual limitations of the human to be im portant in arriving at an optimum design for the vehicle. The report by Ford Motor Company (1973) also identifies important highway targets to be monitored by a driver in his forward and side visual field. These are listed as below.
Forward Field:
-Overhead Signals -Curb Mounted Signals -Curb Mounted Signs -Overhead Signs -Standing Pedestrians -Moving Pedestrians -Fixed Objects (Parked Cars) -C yclists -Intersecting Vehicles -Opposing Traffic -Same Direction Traffic
Side Field:
, -Overtaking Vehicles
The Ford report states that the forward field of view targets in a current model passenger car extend up to 63° on either side of the driver's center line or travel direction. In the visibility model proposed by the Ford Motor Company (1973), emphasis is placed on the use of the peripheral fields of view to fill in the gaps in the driver's visibility of the 360° field surrounding the vehicle. Ac cordingly, it is stated that a driver, when looking at the left rearview mirror, will have his left visual field extended up to a minimum of 70° from the eye po sition (left m irror position in the above example). The value of 70° for the pe ripheral field is based on the experiments conducted by the Ford Motor Com pany, which is described later in this chapter.
Rockwell, et. al„ (1973) developed a computer based tool for evaluating visual field requirements of vehicles in merging and intersecting situations. The model considers road geometry, vehicle design, and driver characteristics to assess the detection potential of visual cues in a merging situation. The authors point out that
"the state of the art knowledge in understanding the role of extrafoveal (or peripheral) vision of drivers for detection of vehicles moving on a side collision course is not suffi ciently advanced and further research in this area is ur gently needed." 51 Visual Fields Affected by Spectacle Frames and Lenses
Smith and Weale (1966) showed that different frame thicknesses can re duce visual field. Behaine and Wick (1976) tested obstruction caused by specta cle frames for 41 pilots. It was found that the obstruction was related to frame size and the largest obstruction observed was about 15° on the temporal side and 8° on the nasal side. It was also observed that a visual angle of 15° above the horizontal meridian was obstructed by the shaft of the spectacle frame. Harper (1966) showed the effect of convergent and divergent lens in producing scotoma and double vision.
Driver Behavior Studies with Interpretations Regarding Visual Performance in the Periphery
Reports of a number of studies found in the driver behavior literature have drawn conclusions regarding drivers' visual performance in his periphery even though the visual field was not among the dependent variables. It will be useful to summarize such conclusions in connection with the present research.
Effect of Alcohol
Belt (1969) found that under the influence of alcohol in open road driving situations, drivers' eye fixation durations increased and the fixation concen trated suggesting a possible narrowing of drivers' visual perception. Also, it was reported that before losing the lateral control of his vehicle, a driver under the influence of alcohol made fixations of short durations over large areas of search. Belt concluded that loss of peripheral vision due to effects of alcohol was responsible for increased foveal fixations on many objects.
Effects of Fatigue
Kaluger and Smith (1970) found that under the effects of sleep depriva tion and prolonged driving, eye movement patterns were less concentrated and occurred closer to the car and to the right of the roadway. This result again suggests the presence of foveal compensation to diminished peripheral detection capacity.
Novice Drivers
Mourant (1970) reported that the novice driver samples foveally a lot more visual details than the experienced drivers. Mourant suggests that the experienced drivers utilized their peripheral vision for information acquisition to a greater extent while even at the end of the driver training period, novice drivers do not reach this stage of visual search. 52 Carbon Monoxide
Recent studies by Rockwell et. al. (1974,1975) show that at higher COHb levels (12% COHb), eye fixation durations tend to increase,, The authors suggest the presence of a narrowing in drivers' visual perception.
Temporal Stress
Bhise (1971) conducted experiments in which drivers closed their eyes voluntarily and drove as much as possible. It was found that upon opening their eyes, drivers tend to sample information primarily through foveal vision.
Foveal and Peripheral Information Processing by Drivers
Of the many investigations to study the role of peripheral vision in driv ing, research by Bhise and Rockwell(l97l) can be considered as most impor tant since it obtained evidence on many hypotheses related to peripheral vision and driving from experiments conducted in the real driving environment.
In one of Bhise’s experiments, test drivers were asked to follow a lead car at a constant headway while staring at a target placed on the rear wind shield of a target car moving in the adjacent lane. The lead car and the target car were moving at constant speed. The angle of separation between the target car and the lead car was varied from 2° to 25°. Based on the results, it was concluded that the test subjects could car follow by monitoring the lead car in extrafoveal vision only.
In a second experiment, the test drivers performed the same task as the first, but the lead car velocity varied in a random manner. Eccentricity angles of 5° and 11° were used. The results of this experiment showed that subjects made quick fixations on or around the lead car whenever the headway distance decreased. This result indicates that the drivers obtain foveal con firmation whenever uncertainty regarding the position and motion of a periph eral target is present.
In a third experiment, Bhise's subjects were required to process infor mation from the target which was stared at in experiments 1 and 2. The task was to identify the signal with a 90° rotation of a Landolt ring kept behind the target car and to maintain a constant headway from a lead car. The separation angles between the two visual targets were separated between 4° and 20°. The fixation direction was not guided by the instructions. The results showed that the performance measures of the two tasks were negatively correlated. Signal detection task was done well when subjects shared higher percentage of time with that task. The car following performance improved as the eccentricity 53 angles decreased. At higher eccentricity angles, subjects cannot stay away from the lead car for a long time.
From the results of the three experiments, Bhise hypothesized that in simultaneously attended foveal and extrafoveal tasks
"the performance in the extra-foveally-attended task would degrade at a faster rate when compared with the degradation in the foveal task as the eccentricity angle increases." Also, "for smaller values of eccentricity angles, performance in the foveally attended task would remain unchanged.11
Peripheral Detection of Vehicle in the Left Adjacent Lane
Ford Motor Company (1973) conducted experiments to determine a prac tical limit for peripheral detection of a vehicle in the left adjacent lane. In one of the experiments conducted at low levels of daylight illumination, subjects seated in a car were required to report the detection of a (static target) car placed randomly in one of the eight angular locations. The car was exposed for approximately one second. The subject’s eyes were fixated on a digital display placed at 25 feet from the subject at the left side m irror position (40° from straight ahead position). The subjects were asked to perform a numerical task to impose a high level of visual attention. Ten subjects were tested with 10 trials at each of the 8 target locations. The results are shown in Figure 15 as the cumulative percent of trials in which the target car was detected as a func tion of the angle of the leading edge of the car from the line of sight. It was concluded that drivers can reliably detect a vehicle in the adjacent lane when it is within 71° from his line of sight during overcast daylight conditions.
In a second experiment conducted at night, the target vehicle was mov ing up slowly with low beam head lamps in the left adjacent lane and was viewed in the background lighting from overhead street lights. The numerical task used in the first experiment was used at the left side m irror position. 40 sub jects were tested in this experiment and the reported results are shown in Figure 15. It was concluded that drivers can reliably detect overtaking vehi cles at night when they are within 73° from the line of sight.
Response to Peripheral Signals under Simulated Driving Conditions
Kochhar (1974, 1973, 1972) developed a simulation of a single lane high way scene for a field of 220° (110° on either side of the center of the highway). The screen was at a distance of 36" from the subject. The film intensity was uniform at 2 candles per sq. ft. The darker areas like trees had an intensity of 1.0 to 1.5 candles per sq. ft. The test subjects were required to drive at sim ulated speeds of 20, 40 and 60 mph and also track the roadway using the steering CUMULATIVE PERCENT OF TRIALS IN WHICH 1 O s o w H ej W E w 0 Q E 40 g S3 w « « W Q H W h h 0 r 100 60 80 20 0 Figure 15: Cumulative Percent of T rials in which Target Vehicle was Detected as a Function of Target of Function a as Detected was Vehicle Target which in rials T of Percent Cumulative 15: Figure
Nighttime conditions with overhead street street overhead with conditions Nighttime lights, Slow moving vehicle as target, as vehicle moving Slow Low beam headlamps, beam Low Line of sight at left m irro r position, r irro m Subjects 40 left at sight of Line 105 ANGLE OF THE OUTERMOST LEADING EDGE OF THE VEHICLE FROM THE FROM VEHICLE THE OF EDGE LEADING OUTERMOST THE OF ANGLE cetiiyAge (eut rm suyb Fr MtrCmay 1973) Company, Motor Ford by study a from (Results Angle, Eccentricity 100 95 IEO SGT (Degrees) SIGHT OF LINE 085 90 Daylight conditions (overcast), (overcast), conditions Daylight angular locations, angular target, as vehicle Static One second exposure at eight eight at exposure second One (40 degrees from straightahead), straightahead), Subjects 10 r from irro m side degrees (40 left at sight of Line 80 75 . J 01 70 55 wheel. In the peripheral task, subjects were required to respond to the detec tion of circular white light signals that were presented at 18 different angular locations. The signals were 3/4M diameter and 16 candles per sq. ft. intensity. The signals were presented in three horizontal planes: horizontal meridian, 26° above and below horizontal meridian0 The eccentricity angles for signals were 38°, 64° and 90° on the left and right side.
In experiment 1, 12 subjects participated. The effects of 1) stimulus location, 2) side of presentation and 3) the demands of continuous central tracking task on the response time to peripheral signals were investigated. The summary of results are as follows:
1. Response times to peripheral signals were found to be normally distributed.
2. Driving speed (film speed or streaming effect) did not significantly affect the response times to peripheral signals. But the response times showed an increasing trend at higher speeds.
3. Response times to signals presented on the right side were shorter than on the left side.
4. The missing of signals occurred mostly when they were presented in the left upper visual field.
5. Response times to signals presented in the far periphery (90 ) were longer when compared with those at 38° and 64°.
Visual Clarity Index
Bhise and Rockwell (1971) reported a methodology for evaluating avail ability of visual information to a driver by studying his eye-movement patterns. This report highlights certain aspects of the role of peripheral vision in driving through a measure called "visual clarity index". The measure is generally de fined as the ratio of the available information to resolvable information under given conditions.
Bhise and Rockwell defined the clarity index for visual acuity consider ing the size of the visual detail as 4. 0 inches.
Index of Clarity, C = a a (ti) 56 where a is the angle subtended by a visual detail of size 4. 0 inches, -n is the eccentricity angle of the visual detail from the visual axis and a (rj) is the visual angle of the smallest detail that could be resolved at eccentricity angle t|. The data for a Ol) was based on data presented by Alpem (1962), as shown below in Table 8.
Table 8
Resolution Angles in the 32° Binocular Visual Field of Drivers with Normal Vision (20/20)
E ccentricity Angle (r|) Resolution Angles D egrees aCn), Minutes
0 0.714 1 1.61 2 2.50 3 2.8 4 3.57 5 4. 00 6 4.55 7 5. 06 8 5.27 16 8.34
The numerator of the index a is dependent on the spatial relationship be tween the visual detail and the visual axis of the driver and the size of the de tail. The denominator a (ti) denotes the visual resolving capabilities of thedriv- ers. If the index is greater than unity then the given visual detail could be dis criminated by drivers.
Bhise and Rockwell developed detailed computer programs considering the above information together with eye movement data from road experiments and the major results can be given as follows:
1) For each fixation location of the eye, the lane markings can be said to be resolved if the value of C ^ 1.0.
2) Percentage of total fixations presenting C ^ l. 0 decreases with increase in preview distance in seconds. « 3) Extrafoveal vision contributes predominantly to the per ception of lane markings than foveal vision. 57 Further work on the Visual Clarity Index is reported by Ford Motor Company (1973). The study made use of the following:
1) Bhise-Rockwell concept of clarity index
2) Minimum visible angles in minutes at eccentricities from 0° to 70°
3) 4. 0 as the clarity index of a 4-inch wide lane marker quite visible at 200 feet from the driver and 2. 0 degrees eccentricity
Using the above information, angular sizes of targets at various eccen tricities that will have a clarity index of 4. 0 were computed as shown in Table 9.
Table 9
Angular Sizes of Objects with a Clarity Index of 4.0 at Different Eccentricities (Ford Study, 1973)
Angular Size of Targets Minimum Visible in Degrees Eccentricity Angles in Minutes with a Clarity Index of 4. 0
0 0.7 0.06
2 1.4 0.12
10 4 .3 0.36
20 9.7 0.81
30 16.8 1. 41
40 25.3 2.13
50 34.6 2.91
60 44.6 3.75
70 56.0 4.70 58 The following general observations can be made with reference to the concept of the Visual Clarity Index discussed on the preceding page. Bhise suggests that the use of visual acuity function in defining the clarity index has many advantages over other functions such as detection and discrimination thresholds. This may be true for the detail (lane markings) chosen for investi gation in the study. If some other detail, such as a passing vehicle in the adja cent lanes is chosen for the investigation of information availability, then the detection threshold may be better suited than acuity in defining the clarity index.
The passing vehicle situation also extends the eccentricity angle to more than 90°. At peripheral angles such as 90°, it is the detection function of the visual system that leads to either subsequent foveal confirmation or other driv ing actions by the driver. Hence, for assessing the information availability in the periphery of driver's vision, the following may be considered as a general form of definition of clarity index:
Clarity 1 _ information available at eccentricity y Index J ------Detection threshold at y
The information may be represented in terms of size, brightness, or contrast of the detail.
It should be pointed out that the angular size of targets reported in Table 9 is only relative to the consideration that a 4 inch wide lane marker at 200 feet from the driver at 2° eccentricity (pavement brightness = 1.0 fL. and contrast = 1.0) is quite visible. For a given target conditions, the clarity in dex at different eccentricities represent a scale of target visibility with the point representing unity corresponding to the threshold visibility.
Field of View as Affected by Drivers' Eye-Head Position Relative to Travel Direction
One of the earliest literature on this topic is by Danielson (1957) and the same has been reported later by Kite and King (1961). Danielson's data regard ing the changes in the uniocular (monocular) visual field as a result of changes in the eye gaze direction and head position is given in Table 10.
The data in Table 10 refer to the left eye. Considering the temporal field, which is relevant to the present research, the following observations can be made: 59 Table 10
Visual Field for the Left Eye
Field Measured Head Eye From the Direction of Gaze Total Rotation Rotation Tem poral Nasal Uniocular Field
0 50° Nasal 100° 10° 110°
0 10° Nasal 100° 50° 150°
0 0° 100° . 60° 160° o -a 0 10° Temporal 1 0 0 ° © 170°
0 50° Temporal 60° 70° 130°
20° Temporal 50° Temporal 60° 70° 130°
60° Temporal 50° Temporal 60° 70° 130°
1) Eye movement up to 50° on the nasal side (50° to the right of straight ahead position) does not affect the left temporal field.
2) A 10° movement of the left eye on its temporal side does not affect the temporal field whereas a movement of 50° on the temporal side reduces the temporal field from 100° to 60°. This would suggest that a large eye movement by drivers to the left of travel direction will reduce the use ful visual field from the eye fixation position on the left side.
3) It can be noted that head rotation to the left of straight ahead position does not affect the left temporal visual field from the direction of gaze. It simply shifts the whole field to the left of travel direction by the amount of head rotation. 60 Danielson reports:
'•Why then do we get the sensation of increasing our peripheral field by turning our eyes to the side when really the fields are less? Apparently it is an illusion because of the relative clarity of the macular and central field which is substituted for the less clear vision of the peripheral field.”
Further information on the effect of eye-head position on the visual field of drivers can be found in literature dealing with ”eyellipse”. The eyellipse is concerned with the determination of the location of drivers' eyes relative to the interior of the vehicle. In one of the eyellipse studies by Devlin and Roe (1968), the effect of eye-head rotation on the visual field has been discussed as follows:
1) With the eyes and the head at zero degree (straight ahead), the binocular field of view in the horizontal meridian is equal to 110°.
2) With 30° of eye rotation, the binocular field reduces to 80°.
3) With 30° eye rotation and 60° head rotation, the binocular field remains at 80°, but the whole field is shifted to the direction of head rotation by the amount of head rotation.
4) Eye rotation of 30° does not produce any change in the temporal fields but affects the nasal field.
5) The temporal fields and the nasal fields are not affected in magnitude by head rotation.
In general, the conclusions of Devlin and Roe (1968) agree with those of Danielson (1957). It should be noted that both reports are not based on obser vations of drivers in actual driving situations. These were manipulation of eyellipse data for straight ahead vision. *
Head Movements during Driving Maneuvers
Robinson et. al. (1971) observed the head movements of 362 drivers who stopped at a T-intersection before making a left or right turn. The results showed that the left turn maneuver was associated with increased search time (head turning time) and increased number of searches (head turns) per subject than the right turn maneuver. On the average, the drivers made 3.7 and 3.5 left and right turns before making a left turn when there was traffic on both sides. The average number of head turns for the corresponding right turn ma neuver when there was traffic on both sides was 2.2 and 0.56. 61 In a second experiment, eight young subjects drove an instrumented vehicle on a four lane highway and performed a lane changing maneuver upon instruction. Data on driver* s head movements during the maneuver was re corded using a special helmet and a potentiometer device. Visual search in term s of average number of head movements was reported as shown in Table 11.
Table 11
Head Movements - Number and Distribution for Lane Change Maneuver (Left and Right) (Robinson 1971)
Lane Change Left Lane Change Right Traffic No Traffic Traffic No Traffic Average Number of Head Turns 2.62 1.88 2.28 1.43
Head Turn Distri bution by Category (Percent) 1) Back (m ore than 33 24 42 56 100°) 2) Side (45° to 90°) 7 16 6 8 3) Left M irror 57 50 3 2 4) Inside Mirror 3 10 49 34 221 Lane Changes (128 Left, 93 Right) Traffic/No Traffic - 76/145
The results show the following:
1. During lane change maneuvers, about 50% of the head turns were directed toward the m irrors0 m 2. The type of m irror sampled is dependent on the type of lane change maneuver. During the left lane change, the left m irror was sampled more frequently and similarly during the right lane change, mostly the inside m irror was sampled more frequently.
3. During lane change maneuvers, about 30 to 40 percent of head turns were greater than 100°. 62 Robinson, in addition to number of head turns, also determined 1) the total time for the maneuver, 2) visual input time (time to look at the mirrors, side and back), and 3) time loss due to eye/head movement (assumed as 300 msec s. from laboratory studies). Using the temporal analysis, it was con cluded that approximately 30 percent of the maneuver time was spent by drivers to acquire necessary information for making a lane change maneuver and about 15 percent of the time is lost due to eye/head movement.
Robinson's head movement data provide only gross values for different categories of movements. It does not provide a distribution of head turns with in the m irror location angle. The visual search time analysis is based on the assumption that the presence of head turns indicates a search. But it is pos sible for search to occur with only eye movement and no head movement. Hence, the data on visual search time provided by Robinson may be less than the actual search time.
Attention in Driving
Connolly (1968) and Schmidt (1968) have discussed some of the problems related to attention and peripheral vision of drivers. Attention is considered to be an important variable affecting visual perception. Perception of vehicles and other targets which are on a collision path is delayed when drivers are attend ing to other targets. Visual demands in driving vary depending on driving con ditions and may be represented by the amount of eye movements required to process information. In urban driving, the eye movements vary from 45 to 60 degrees on either side from the center, in suburban driving, eye movements vary from 30 to 50 degrees and in rural, they vary from 10 to 15 degrees.
Bhise (1971) expresses the level of attention in a given driving situation by the amount of time a driver can share with other targets such as signs. The time sharing index in Bhise's studies was given by T./T used where Tj is the total time available for sign reading and T used is the time spent in reading signs. Increase in traffic density or close car collowing decreased T used and hence increased the value of the time sharing index T^/T used. Rockwell (1972) reported that during normal driving, drivers' eye fixations are located in the center of the road near the horizon spreading over a 6° x 3° zone. In many re search investigations at Ohio State University, eye movement performance mea sures were related to this area. (Example, Mean Time Spent in this Region, Mean Time away from this Region, Percent of Total Trial Time in this Region.)
In general, active visual search is considered to be an indication of high driving demands. Peripheral Motion Perception Drivers 63
Perception of motion is a special case of direction and distance percep tion. It involves alterations of the retinal image in one of the following manners:
1. The image may move across the retina.
2. The size of the image may either grow or shrink.
In driving, the size of the image seldom remains constant. It can be seen from the instrumentation chapter that the visual angle subtended by a passing vehicle is maximum when it is at 90° from the eye position and reduces gradually in either direction. Therefore, the nature of retinal image alterations involves both movement and size changes. In driving, the dynamics of the eye, objects on the road, and the environment are involved in producing the alterations in the retinal image. The environment will move and produce a streaming effect only if the vehicle is in motion. Table 12 shows the type of motion perception (foveal or peripheral) for different combinations of eye position, motion of ob jects, and environment. The following can be summarized as general results in the motion perception. i i •A driver cannot turn his eyes much farther than 10 degrees left or right and still perceive meaningful motion of the car ahead.
• Minimal Perceptible Speed 1. 1 to 2 minutes of arc/sec in the presence of stationary reference objects and when the eye is fixating on the target
2. 15-30 minutes of arc/second with no stationary ref erence objects (in dark)
• The threshold of speed and the threshold of displacement are lower in photopic visioii than in scotopic vision.
• The higher the contrast of the moving object, the lower will be the threshold of motion.
• The perception of motion is more accurate with increased exposure time.
•Perception of motion is better in the fovea than in the periphery. 64 Table 12
■type of Motion Perception by Eye Position and Motion of Objects and Environment
Type of Eye Position Objects Environment Motion Perception
1) Fixed Moving Stationary Peripheral (Vehicle Stationary)
2) Follow the ob jects (Pursuit Moving Stationary Foveal Movement)
3) Fixed Stationary Moving P erip heral
4) Fixed Moving Moving P erip heral
5) Follow the objects Stationary Moving Foveal
6) Follow the objects Moving Moving Foveal
• Motion is overestimated when perceived peripherally.
• Motion in depth is encountered most frequently in driving. Observation time required to recognize motion in depth decreases with increasing speed. « • The driver recognizes the movement of the car from the apparent flow or streaming of the objects in the visual field in the opposite direction.
• Dynamic visual acuity suffers when the pursuit eye move ments are not capable of holding a steady image of the target on the retina. r • Blurring of image decreases its contrast. • Smooth lateral pursuit eye movements are possible up to 65 a velocity of 30 degrees per second.
* * Dynamic visual acuity is highest in the fovea and less in peripheral vision.
Most of the results on the perception of motion are related to relatively small size stimuli and according to Graham (1965),
" large moving stimuli need not be expected to give effects like those obtained with small stimuli. We know little about such effects from a quantitative point of view, A program designed to improve our knowledge in this re gard seems desirable.”
Current Research
Currently, research supported by the National Highway Traffic Safety Administration is underway in the field of peripheral vision and driving in two areas. The Systems Research Group at the Ohio State University is investi gating the ’'Utility of Peripheral Vision to Motor Vehicle Drivers”. Specific objectives of this research are:
1. to determine the magnitude of head turns associated with m irror samplings by drivers during lane changing and driving maneuvers on the freeway.
2. to determine the peripheral detection angle of drivers as a function of foveal task loads, passing vehicle relative velocity, and eye/head position in the laboratory.
3. to validate the results obtained in the laboratory in the real driving environment.
Head movement data in the above study was collected using 50 test sub jects (both young and aged). The data was gathered in the subject's car using portable instrumentation without the knowledge of the drivers. The subjects performed maneuvers such as lane changing and merging which increased the m irror samplings. Preliminary analysis of head movements data show the fol lowing results. (Kretovics, 1976)
1. Aged subjects make larger head turns than young subjects. i 2. Percent of m irror angle by which subjects turn their heads during m irror sampling is found to be as follows: Head Turn
(Mean Percent of M irror Angle) Young _____ Aged
Left Mirror 61% 70%
Inside Mirror 44% 56%
The m irror locations in the vehicles of test subjects were found to be distrib uted as follows:
Angle of M irror Location in Degrees from Straight Ahead
Mean 50th Percentile 90th Percentile
Left Mirror 57.5 57.0 72
Inside Mirror 44.4 43.0 52
Also, currently in California, Human Factors Research Incorporated is investi gating the development of training techniques to improve peripheral vision of d riv ers.
Summary of Literature Review
Based on the literature reviewed, the following conclusions can be made regarding visual field requirements of drivers and their peripheral visual capa bilities.
Visual Field Requirements:
1. In driver licensing there is no uniform standard for testing drivers1 visual field limits as well as a uniform criteria for the issuing of a license. 2. No useful information is obtained from the conclusion that limited visual fields and drivers' accident experi ence are not correlated. To understand the effect of limited visual fields on driving behavior, intermediate criteria not related to accidents should be developed and used in the evaluation of visual fields of drivers.
3. Automotive designers have developed standards for field of view from automobiles in which it is assumed that drivers' peripheral visual field is assumed to fill in the gaps in the 360° field around the vehicle. In one of the such standards, drivers' peripheral detection field has been taken as 7 0°.
Peripheral Vision of Drivers
1. From Bhise's studies, it can be concluded that
i) drivers can monitor driving (primary task) extra- foveally under no uncertain conditions.
ii) with the introduction of uncertainty in the primary task, frequent foveal confirmations are necessary to monitor driving extrafoveally0
iii) with the introduction of a peripheral information pro cessing task, performance in both the tasks deter iorate at higher eccentricity angles. Performance degradation in the peripheral task would be at a faster rate than in the foveal task. At smaller eccentricity angles, the foveal task performance would remain unchanged.
2. The results of the experiments by the Ford Motor Company (that drivers can detect passing vehicles when they are within 70° from eye position) should be verified because the experiment used stationary targets and stationary subjects. Other than the re port by the Ford Motor Company, there seems to be no research related to the detection of passing vehi cles from behindo
3. Spectacle frames with thick shafts may obstruct the view of objects up to a 15° visual angle. 4. Drivers under stress (fatigue, alcohol and carbon mon oxide) may show reduced efficiency in their peripheral vision.
50 Since the use of peripheral vision while driving seems to be lacking with novice drivers, the driver education programs may explore the possibility of providing train ing in peripheral vision to young drivers.
6. No experimental evidence exists regarding the peripheral detection capabilities at different target contrasts and background luminance conditions such as dawn, dusk, etc.
7. Bhise's Visual Clarity Index concept combines the infor mation available and the resolvable information. This concept may be extended to the detection of a passing vehicle situation. The clarity index can be used in com puter models (Bhise, Rockwell, 1973) designed to study the visual field requirements in any driving situation.
8« Head movements during visual search simply shift the drivers' field of view by the amount of head turn. It does not affect the limit of the peripheral field of the d riv ers.
9. Small eye movements up to 10° on the temporal side of any eye will not affect the corresponding eye's temporal visual field. But eye movements of 50° magnitude to the left will affect the left temporal field greatly0
10. In freeway driving during lane change maneuvers, about 50% of the head turns were directed towards m irrors. About 30 to 40 percent of head turns were greater than 100°. It is stated experimentally that about 300 msecs. is spent in eye/head dynamics. Hence, based on the number of head turns occurring within each maneuver, it has been reported that about 15 percent of the ma neuver time was lost due to eye/head dynamics.
11. In driving, the higher percentage of drivers' eye fixations lie in a central 6° x 3° area. Increased eye movement activity in this area may be considered as a result of in creased driving task demands. 12. Very little experimental evidence is available regard- 69 ing peripheral detection of large moving targets such as vehicles.
13, In driving, changes with eye fixation produce changes in the foveal task and peripheral targets and/or char acteristics. Driving research thus far has thrown light on the visual search behavior of drivers. Exten sive literature on human visual performance may help to model drivers' information acquisition capabilities and limitations at the fovea. But very little is known about drivers' peripheral vision capabilities. Any in formation made available in this area may be useful to automotive designers, highway construction engi neers, and driver licensing and highway safety author ities. BLANK PAGE CHAPTER 3
THE RESEARCH PROBLEM AND METHODOLOGY
Research Problem
The overall objective of this research was to investigate some aspects of human peripheral vision with reference to the driving task. It was discussed earlier in Chapter 2 that the field of vision of drivers rotates within the 360° horizontal field as the eye fixation point changes. To a greater extent, the changes in the eye position are affected by the driving maneuver, traffic conditions, and purpose of the trip. Associated with each maneuver and eye position are a number of relevant peripheral targets which the driver may be required to detect for purposes of safe driving. These maneuvers, such as normal driving with the eyes in the straight ahead position, sign reading, and mirror sampling will form the scenario for testing driver’s peripheral detection capabilities, it was decided to use the simple driving scenario, i.e ., normal driving with the eyes in the straight ahead position.
For the above scenario of driving as indicated in Chapter 2, a number of relevant peripheral targets can be identified. These are:
1. passing vehicles in the left and right adjacent lanes from behind
2. vehicles merging from behind
3. vehicles moving on the crossroad
For purposes of this research, vehicles passing from behind in the adjacent lanes were considered as peripheral targets.
Generally, responses to peripheral targets may be of the following form :
71 72 General Response Response Related to Driving (Scenario) Detection Awareness of the presence of vehicles passing from behind.
Recognition or Awareness of the nature of targets Discrimination such as trucks, cars, motorcycles, etc.
Judgment of Awareness of the lane position of D istance target vehicles.
Judgment of Motion Awareness of motion itself, change in motion, and direction of motion.
For this research, a simple detection response was considered for the subjects. Hence, the performance measure is simply the peripheral detection angle. The major variables selected for research were as follows:
Independent Variables
Foveal Task Demands:
Literature review presented in Chapters 1 and 2 indicate that the level of attention at the fovea will be one of the primary factors in determining the extent of drivers’ visual field. Also, in the previous laboratory studies, a central tracking task was used to obtain subjects' attention. But in this research it was proposed to keep the foveal task to one of visual search only. This decision was made to relieve the psychomotor mechanisms from inter fering with visual processes. While it is true that in driving, psychomotor processes are combined with visual processes, it should be remembered that systematic knowledge about the functioning of the visual system should first be studied without confounding factors other than visual in nature.
Target Contrast:
Among the target variables, target contrast was considered for investigation. It was shown in Chapter 1 that contrast sensitivity decreases with decreased background luminance. Hence, a worst case situation will be to detect targets of low contrast in the far periphery at low levels of illumi nance such as dawn, dusk and nighttime conditions. Therefore, it was decided to test the peripheral detection angle under subthreshold, threshold, and superthreshold conditions. Controlled Variables 73 Target Size:
The target size variable was decided to be controlled at a value that is comparable to normal driving conditions. From data on visual angle subtended by highway vehicles given in Chapter 4, it can be seen that passenger cars passing in adjacent lanes subtend about 34° at 90° from the travel direction. Hence, the target sizes found in a highway environment during a passing situation will be much greater than the target sizes used in the past laboratory experiments. Ricco's law regarding spatial summation holds good only up to a target size of . 10 minutes of arc in diameter. Hence, varying target size may not produce useful information. Under conditions of no spatial summation, target luminance or contrast alone is responsible for detection. Hence, earlier decisions to use target contrast as an independent variable in the research is supported.
Target Motion:
Since the objectives of research was to investigate the peripheral detection angles for passing vehicles, target motion gets introduced in the experiment automatically. A low relative velocity on the order of 0.5 mph or less was decided for the research and this level was to be maintained through out the research. This corresponds to an angular speed of about 5°/sec. The reason for selecting a low angular speed for the target was:
1. low speed is the next logical case to stationary target and
2. higher relative velocities in practice if involved in a critical situation do not provide sufficient time for the driver to act.
Research Methodology:
The basic research methodology was to simulate the highway scene with moving targets in the laboratory. It was decided to provide a foveal display task at the center of the simulated road scene. Hence, the subjects were to attend to a foveal display task and a peripheral task.
The nature of the foveal task was to simulate different levels of search activity on the part of the subjects. Because of the dual task situation the subjects were to give primary importance to foveal task compared to peripheral task. The level of demands from the foveal task were to be maintained less than or equal to some critical load for each subject so that the subjects could —"maintain a given level of performance in the foveal task. 74 The peripheral task was to report the first detection of a vehicle moving from behind. Hence, the performance measure of interest was the peripheral detection angle from the eye position. Subjects vary in their detection threshold with respect to target contrast. Hence, different levels of target contrast or visibility to be used in the research were to be based on each subjects' threshold contrast for detection.
The following were some of the specific considerations in the research methodology:
Subjects: The test subjects were to be given a complete eye examination to screen for eye defects.
Age: . It was decided to test subjects from both young <20 to 35 years) and aged (above 50 years) groups.
Baseline Data: It was decided to collect data on subjects' normal visual field and simple response time for purposes of correlating with the research re s u lts .
Critical Foveal Load: It was decided to determine the critical level of loading for each subject on the foveal task. Other levels of foveal task were to be based on this critical load.
Threshold for Target Visibility: It was decided to use the target (vehicle) image in a stationary condition at 90° eccentricity and determine each subjects' threshold for visibility (contrast) of the target. This data was to form the basis for the selection of different levels of target contrast or visibility.
Streaming Effect: Since during driving highway objects stream across the visual field, it was decided to simulate the streaming effect in the laboratory. Visual Scene of Driving: To produce the visual scene for target presentation the following were decided:
1) make movie films of a freeway from a moving vehicle to cover more than 100° of left visual scene from the straight ahead position.
2) left field was selected for testing since vehicle pass normally on the left side.
3) These films were to be projected in the laboratory on screens subtending the same angle which was covered in the films.
Peripheral Target Presentation: Peripheral targets in the form of passing vehicles were to be simulated by superposing a vehicle image using a slide projector on the projected screen.
Target Motion: It was decided to rotate the slide projector at desired angular speed to produce motion of the projected image on the screen.
Data Collection: It was decided to interface the laboratory set up with the PDP-8/L Computer for purposes of foveal task control and recording subjects’ response in the foveal and peripheral task.
The next chapter describes in detail the laboratory arrangement for exper imentat ion. BLANK PAGE CHAPTER 4
LABORATORY SET UP AND INSTRUMENTATION
Introduction
This chapter describes the different aspects of the laboratory set up and instrumentation in the following sections:
1) Laboratory Set Up 2) Simulation of the Driving Scene • 3) Simulation of Peripheral Targets 4) Control of Target Size, Lane Position, and Relative Velocity i) Visual Angles Subtended by Passing Highway Vehicles i i ) Simulated Target Images i i i ) Relative Velocity and Angular Speed of Highway Targets 5) Brightness Contrast of Targets i ) Target Visibility i i ) Average Luminance of the Target and the Background i i i ) Target Visibility Values Comparable to Highway Target Contrasts 6) Foveal Task 7) Subject's Response Console 8) Experimenter’s Console 9) Computer Laboratory Set Up Interfacing
Laboratory Set Up Considerations
The following were the primary considerations in setting up the labora tory for the research.
1. To simulate the visual scene of freeway driving of at least 100 visual degrees from the point of eye fixation
2. To introduce peripheral targets as passing cars from behind the d riv er
3. To control the size of the cars and thereby simulate the lane position of the cars
4. To control the angular speed of the targets and thereby simulate a desired relative velocity of cars 77 5. To control the brightness contrast of the target vehicles and thereby simulate the brightness contrast of highway vehicles observed inreal driving situations
6. To develop a task at the foveal position of the eye which can be used to load the driver at desired levels and to simulate the intense foveal visual search patterns of drivers during the laboratory task
Simulation of the Driving Scene
The basic approach taken was to film the highway scenes from a moving car and to project these films in the laboratory to produce the required visual scene. Films were taken using a 16 mm movie camera fitted with an 8 mm lens. The 8 mm lens covered approximately 65° horizontal field of viewa The camera was placed at two different positions: 1) facing the travel direction (straight ahead) and 2) at 70° from the straight ahead position facing the left scene. The films were made in the Interstate 270 freeway with three lanes of traffic in either direction. The general illuminance at the meter on the point of observation was approxim ately 8, 000 foot candles (m easured by GOSSEN LUNASIX Electronic Exposure Meter). The films were shot in a section with no significant objects such as bridges, overhead signs, etc. in the scene. The camera car was always in the slow speed lane (far right) moving approximately at a speed of 40 mph. The film was shot at a film speed of 16 frames per second.
In the laboratory, the two films were projected by two 16 mm projectors (Bell and Howell Model 173 BD) on two 92" x 66" screens. The arrangement of the screens, projectors and the position of the subject seat are given in Fig ure 16. By placing the subject’s seat at 72" from the center of both the screens, each screen provided a horizontal field of view of 65°. Between the screens a gap of 5 degrees was provided and covered by black hardboard. The presence of this gap made it possible to project the front scene and the left scene with out being synchronized. This was also later found to be unnoticed by test sub jects. Hence the expensive hardware for synchronization of the two scenes was avoided.
The two projectors were modified so that the output of the projector lamps can be controlled independently by two rotary rheostats without affecting the projection speed. The projectors also had the capability of varying the film speed from 800 to 1000 frames per minutea The projectors were located on the right and behind the subject at a distance of 124" from the center of the screen. They were raised about 77" above the ground to avoid any shadow of the subject’s head on the screens. The projectors were fitted with 0.5 inch, f 1 . 4 wide angle lenses. The projection of the scene films were made contin uous by using film loops in the specially fabricated loop film cartridges. 79
Line of Sight and [ Major Axis of Head \ (-32. 5°) 1 Foveal I Display
/ Front Screen
< -7 0 °K .
(-90°) Slide P rojector Subject j 16mm (-102.5C Left Screen Projector
16 mm - t j } - P rojector I
Figure 16: Schematic'Diagram of the Laboratory Set up for Screens and Projectors Simulations of Peripheral Targets 80
In this research, movement of highway vehicles from behind the test subject was simulated by the superposition of the images from a slide projector on the left scene. The location of the slide projector was to the right of the subject at 90° as shown in Figure 16. It was located at a distance of 60 inches from the subject and approximately 80" above the ground so that the projection beam passed above the subject's head without producing any shadow on the left screen.
The slide projector used was "Kodak EKTAGRAPHIC RA 960 Random Access Projector". It was fitted with 2 inch, f 3.5 wide angle projection lens, and a 500 watt projection lamp. The projector was modified for independent control of lamp output through a rotary rheostat.
The slides used for targets contained the image of the automobiles against a black background so that when projected only the image of the target vehicles was superimposed over the left scene. The size, color and nature of targets used in the research are explained in another section of the report.
The motion of the projected images was obtained by placing the slide projector over a rotating turntable powered by a 1/15 HP "Dayton" AC/DC Gear Motor. The speed of the motor could be varied from zero to 9 rpm through a speed control device.
Introduction of peripheral targets was restricted to the testing region of the visual field, i0 e. from about 115° to about 50° in the inward direction. The slide projector rotation was made to reverse in both directions when the limits of test region was reached by means of two microswitches which operated the reversing mechanism of the motor. The turntable had provisions for presetting the test region to any desired value.
Control of Target Size, Lane Position and Relative Velocity
One of the considerations in the laboratory simulation of highway targets on the screen was to produce vehicle images on the screen that were compatible with the scene that was projected on the left screen. Chie method of achieving the realism of vehicle size, location, and relative velocity is to film such tar gets in the left scene and project that film in the laboratory on the left screen. Preliminary efforts in this direction encountered the following problems:
1. The general lighting level during complete filming should be constant for the complete length of the film. This was made difficult due to constant changing of the daylight conditions. 2. Bright overcast day is preferable to eliminate leading ° j. shadows of target vehicles. Again this could not be achieved during filming and the target shadows when present acted as a predominant visual cue.
3. Maintenance of constant low relative velocity for targets was difficult in the normal traffic situations.
4. By this method, using target relative velocity as an inde pendent variable in the research was found to be uneconomi cal for it leads to high cost of filmmaking since each target has to be filmed at each of the levels of relative velocities. On the other hand, it was possible to change the relative speed of targets in the laboratory during projection by sim ply changing the film speed. But this was accompanied by a change in the speed of the scene simulating a change in the speed of the subject's vehicle.
The following sections describe how some of the above difficulties were overcome by the superpositioning of the vehicle images on a moving highway scene. The discussions also relate to the achieving of the realistic target sizes, and relative velocities in the simulation.
Visual Angles Subtended by Passing Highway Vehicles
The visual angles subtended by passing highway vehicles at the driver’s eye depends on l) longitudinal and lateral separation between vehicles, and 2) the size of the passing vehicle. It is reasonable to assume that the leading edge would be the first to stimulate the receptors at the retina. But the largest visual angle subtended is due to the overall height of the car. Let 36 and 54 inches be assumed as representative values for the height of the leading edge and the body of the car respectively. For the case of vehicles passing on the left two lanes, let the lateral distance from the driver be assumed to be 7 and 19 feet. Let the eye height of the driver be assumed to be 44 inches above the road surface. Then the visual angles subtended by the leading edge and the body of the passing vehicles can be calculated and given approximately by the values in Table 13.
Let the test region for the visual field be assumed to range from 70° to 100° from the straight ahead position of the eye. In this range, it can be seen from Table 13 that the visual angles subtended remain approximately the same. But upon further calculations it can be seen that these angles decrease markedly as the angular position of the target vehicles approaches zero degree. Angular Position of the Passing Vehicle In Degrees from Travel Direction NOTE: Driver's eye height = 44"=height eye Driver's NOTE: 100 110 50 60 90 80 40 70 Lateral distance between vehicles = 7' and 19' in lanes 12 and inlanes 19' and 7' = vehicles between distance Lateral Freeway lane 12'width =lane Freeway Visual Angles Subtended by Vehicles Vehicles by Subtended Angles Visual on the left on the left onthe left the on is ae Second Lane Lane First 21.92 22.21 21.08 21.92 21.08 76 .727.07 6.87 17.69 15.10 19.68 Leading EdgeLeading (36") nteAjcn Lanes Adjacent the in Visual Angles in Degrees Subtended By:inDegrees Angles Visual Table 13 Table 8.39 .833.96 8.78 .932.58 8.39 .134.43 8.91 .833.96 8.78 .822.97 5.78 7.74 is ae Second Lane Lane First on the left on the left on the left on the 32.58 30.28 Body of the Vehicle (54")VehicleBody of the 13.43 12.64 13.23 10.33 13.23 12.64 11.65 8.68 82 83 Simulated Target Images
In the simulation one should not try to achieve the values of visual angles subtended as shown in Table 13 because the target vehicle sizes in the simulated scene will depend on the screen size and the subject’s distancefrom the screens. Hence, the approach in the laboratory should be to achieve target sizes that are compatible to the size of the scene that is projected. An empirical method of achieving this compatibility is to project a left scene film containing highway vehicles of known size in the simulation set up and measure the size of the ve hicle image produced on the screen. For example, a vehicle with an actual height of 54 inches was observed to have compatible sizes of 36" and 20” when moving on one and two lanes respectively to the left of the subject's car at 90° angular position. Using this information together with the distance of the slide projector from the screen, the focal length of the projection lens and the frame size of the slide, the size of the vehicle image on the slide can be determined. For the given arrangement of the screens, slide projector and a target image size of S inches at 90° angular location, the image height (s) on the on the slide can be calculated as shown in Figure 17. The figure gives a view along a plane passing through the slide projector, subject, and the left screen and parallel to the front screen. | I Simulation of the lane position of the car can be done by making the car image in the slide smaller corresponding to the compatible size of the car in the second lane and also suitably positioning the image above the bottom edge of the slide frame.
Relative Velocity and Angular Speed of Highway Targets
One of the target variables affecting peripheral detection of a vehicle passing from behind is its relative velocity. The relative velocity is the dif ference in speed (MPH) between the two vehicles. Since the paths of motion for the two vehicles are parallel, a constant relative velocity would produce dif ferent target angular velocities (degrees/second) at different angular locations of the target. {See Figures 18 and 19.) As described earlier, the motion of the target image in the laboratory was produced by rotating the slide projector at approximately constant speed. It can be seen from Figures 18 and 19 that for a given constant angular speed, the relative speed of the passing vehicle would increase as the vehicle moves inward from a 90° position. It can also be noted that angular speeds less than 10 degrees per second will represent approxi mately constant relative velocities for targets moving inward up to 60° angular location. 84
P ro je c to r L eft w ith Screen 2 " len s
1 5 0 " . ^
\ i Subject1 s Eye 80
1 r 44"
Bottom of S creen
76. 0" 59
(Figure not to scale)
s = 0. 0137 S inches where S is the target image size on the screen in inches.
s is the target image size in the slide in inches.
NOTE: The size refers to the vertical or height dimension of the target.
Figure 17: Diagram Showing the Relationship of Scene Image Size to Slide Image Size 50
'S ta 40 — ■3o 73 u ■a o a> CO 0) ft ©GO a> 30 - (joa) Q tn
4. 0 mph 0 *3d) a) a. CO n 3.0 mph cd 1 < 2.0 mph 10 —
3 .0 mph 0.8 mph 0.5 mph 0.3 mph
9080 70 60 50 % Target Location in Degrees from Straight Ahead Figure 18: Angular Speed of Vehicles Passing in the Left Adjacent Lane at Different Angular Locations for Different Constant Relative Velocities 86
■8 Is ao ■a u ■a uo 0> CO Q> O. CO Q> 0) fl>fell Q a 00 20 H bo u H 0 15 H ’Sa> cu CO s 10 -J 5. 0 mph 1 a < 4. 0 mph 3. 0 mph
2. 0 mph 1. 0 mph 0.5 mph T
90 80 70• 60 50 Target Location in Degrees from Straight Ahead Figure 19: Angular Speed of Vehicles passing in the Left Second Lane at Different Angular Locations for Different Constant Relative Velocities Brightness Contrast of Targets 87
The brightness of the scene or targets in the simulation set up depends on 1) the output flux of the projection lamps, 2) the transmittance of the film and 3) the reflectance properties of the screen. Since the maximum output of the projection lamps is limited to 750w it is practically not feasible to create the daylight luminance conditions of the target and background in the laboratory. But it is possible to accurately report the average luminance conditions of the scene and the target image adopted in the research. Also it is possible to cre ate target brightness contrast levels that are comparable to those found in real driving conditions. In general, the absolute contrast of any target can be given by
P* ' 0 ' where Bj. and B^ are the luminance of the target and background.
Before describing the procedure of measuring the target and scene lumi nance, the following limitations of the simulation set up have to be noted:
1. The maximum luminance conditions of target and scene were limited by the wattage of the projection lamp of the slide pro jector and film projector.
2. Slides of dark cars did not produce a uniform image over the bright scene; instead it produced a lot of bright highlights which were seen against dark background. But slides of light (white) cars produced a much better uniform field than dark cars. Hence, the laboratory set up could handle only light cars in a satisfactory manner to simulate highway targets.
Target Visibility
The term 'target contra'st' in the simulation situation may not be appro priate because of the nonuniform field produced by the target images from dif ferent slides. Because of the nonuniformity in the field, each subject may be stimulated by different highlights of the image on the screen. Hence, accord ing to Fry (1976) it is appropriate to use the term 'target visibility' of a given target by measuring the wattages of the projectors producing the background and the target. For a given wattage of the background and for a given periph eral angle, the threshold of the target visibility can be measured by varying the wattage for the target. The wattages cab then be translated into values for the average luminance for the target and the background. 88 Average Luminance of the Target and the Background
The following procedure was adopted to measure the average luminance of the target and the background. (See Figure 20.)
• Consider the case of the target.
•A short focus lens (+20) was placed in front of the projection lens of the slide projector.
•A clear focused image of the target vehicle from the slide was picked up on a screen at a close distance to the projector.
• A rectangular hole covering just the image vehicle was cut in the screen.
• To measure the total flux coming through the hole, 'SPECTRA' Photometer Model FC 200 was used.
• The probe of the photometer contained a solid-state selenium photovoltaic cell. This cell was placed directly in the path of the flux coming through the hole.
•A neutral density filter 4. 0 had to be used between the cell and the screen hole to read the photometer within the avail able ranges.
• The wattage of the projector was set at its maximum by set ting the variac knob at 100. Then the photometer reading, which represented the flux in relative units was noted.
• The wattage of the lamp was reduced in steps by setting the variac knob at 90,80, 70, „ . . and 0 and every time the photometer reading was noted.
• The above procedure was repeated two times to check the accuracy of the readings.
• The above procedure was repeated without the target slide in the projector.
• Thus, the flux corresponding to the target area with and with out the slide for different wattages were recorded. Screen with a rectangular hole just to cover the target Image. Screen painted black on the photocell side.
Focused Image of the Target
N. D. Filter' / 4.0 Projection Lens Short Focal Length Lens Target Slide (+20)
Probe of the Photo Figure 20: Schematic Diagram of the Set-Up for meter with its photo Light Flux Measurement cell facing the Image and covering the com plete hole.
oo to 90 ° From the data on flux, the slide transmittance Tg for a given variac setting i was calculated as follows: * (Flux with Slide). T (Flux without Slide)|
•Proper measurements should yield the same transmittance values at different variac settings.
• Then the short focal length lens and the associated screen in front of the projection lens was removed.
•The left screen (on the subject’s left side) was illuminated by the slide projector without the slide.
• The luminance of the left screen was then measured by the photometer using its probe with the photogrid. The probe was placed facing the screen at 32" away from the screen and 32" above the ground level. The photogrid was capable of averaging the luminance over 60°.
• The average luminance measurements in foot Lamberts were made at different wattages of the lamp (variac settings). Let this be called Ls^.
• Then the average luminance of the vehicle image (B^) pro duced by the slide on the screen was calculated as follows:
where i refers to the variac setting.
The above procedure was followed to determine the transmittance of the left scene film and the average luminance of the left scene (B^). During flux measurements with the film, care was taken to record the readings after they stabilized. Figures 21 and 22 show the variation of the flux from the slide projector and film projector respectively at different variac settings. Figures 23 and 24 give the luminance for the target and background scene. From the average luminance values for the target and background, the target visibility for different wattages (variac settings) can be computed as follows:
Target visibility ) _ + ®bj) - Eb* _ Ety at setting i Figure 21: Luminous Flux Produced by the Slide P ro jecto r with and without and with r jecto ro P Slide the by Produced Flux Luminous 21: Figure Luminous Flux in Relative Units 100.00 50. 00 50. 00 H 10.00 .0 H 5.00 1.00 H 1.00 0.50 0.10— 0.05 h ld tDfeet aic Settings Variac Different at Slide the 40 Variac Setting in Divisions in Setting Variac vrg ld Tas tac = 11 = ittance Transm Slide Average 60 80 100 With Slide Without Slide Luminous Flux in Relative Units 10.0 0.1 0.5 1.0 .0 5 Figure 22: Luminous Flux Produced by the Left Scene P rojector rojector P Scene Left the by Produced Flux Luminous 22: Figure ihadwtotte im tDifrn Vra Settings Variac ifferent D at Film the without and with 40 Variac Setting in Divisions in Setting Variac vrg Fl as tac = 9.41% = ittance ransm T Film Average 60 80 100 Without With Film Film Luminance in Foot-Lam berts 10.0 0.01 .5 - 0.05 0.1 0.5 1.0 .0 5 Figure 23: Left Screen Luminance without Slide and with Slide (Target (Target Slide with and Slide without Luminance Screen Left 23: Figure mg)atDfeet aic Settings Variac Different t a Image) 40 Variac Setting,in Divisions Setting,in Variac 8060 100 ^ With With ^ ' Without Without ' (Calculated) Measured) (M Slide Slide Slide Luminance in Foot-Lam berts 0 -i .0 0 1 0.01 - .0 5 05- 5 .0 0 0.1 . - 0.5 1.0 iue2: etSeeLmnnewtot im n ih im (Back Film with and Film without Luminance Scene Left 24: Figure - - rud cn)a ifrn Vra Settings Variac Different at Scene) ground 060 40 Variac Setting in Divisions in Setting Variac « 80 100 ^ With With ^ (Calculated) „ Without Without „ (Measured) Film Film 95 At this point it should be noted that the above target visibility C is the same as the target contrast with uniform fields of target and background luminance. For reasons explained earlier, the term target visibility is preferred instead of tar get contrast.
Target Visibility Values Comparable to Highway Target Contrasts
If the average background scene luminance is kept constant correspond ing to the full wattage of the film projector, then for different wattages of the slide projector the target visibility can be given by the curve in Figure 25. (Also see Table 14.)
Bhise (1976) measured the brightness of approximately 250 oncoming vehicles and their backgrounds and reported data on the background luminance, target luminance and the absolute contrast as shown in Table 15.
To get further insights into the ranges of background luminances and target contrasts existing in the real highway environment, the following efforts were made. The background luminance and the luminances of vehicles parked in a parking lot (paved with asphalt) were measured with a photometer (SPEC TRA" Model FC 200). The measurements were made during daylight (4:00 PM) and dusk (8:30 PM) conditions. Table 16 gives the data on the measurements including the vehicle contrast.
By comparing the values of luminances from Tables 14, 15, and 16, it can be seen that there is a large difference between the target and background luminances obtained in the laboratory and those in the real world. But the con trast levels in the laboratory can be developed to equal those in the real world situations.
Foveal Attentional Task
Purpose:
The purpose of the fove&l task was to require the subject to direct his attention to the demands of the foveal task during the testing of his peripheral visual performances. By varying the demands of the foveal task, it was ex pected to vary the subj ectr s attentional level at his fovea.
Device:
The device fabricated for the above (see Figure 26) was a visual display consisting of six lighted digits arranged in two rows and three columns. The horizontal and vertical angles subtended by the display at the subject's eye seated six feet from the display were respectively 6 and 3 degrees. Luminance in Foot-Lam berts 10.0 0 0.05 0.1 . - 0.5 - .0 5 iue2: agtAvrg uiac n Tagt iiiiy at Visibility arget T and Luminance verage A Target 25: Figure 1.0 . 01 - — I I I I I I 1 0 0 80 60 40 re Visibility arget T Variac Setting in Divisions in Setting Variac ifrn Vra Settings Variac Different T arget A verage verage A arget T Luminance r “ T 100 — 0.5 - 0.05 “ 10.0 0.01 0.1 .0 5 1.0 96 Table 14'
Target Visibility for Different Wattages of Slide Projector
Average Target Target Visibility Variac Luminance i Setting Bj. ft. Lamberts ^ 1 0 0
100 1.08 7.0634
90 0.831 5.4349
80 0.5706 3.7318
70 0. 3623 2.3695
60 0.2282 1.4924
50 0.116 0. 7586
40 0. 0448 0. 2930
Bk = 0.1529 ft. Lamberts D100
Table 15
Data Absolute Contrast of Highway Vehicles by Bhise (1976)
Background T arget Absolute Luminance (ft. L .) Luminance (ft. L .) Contrast P ercentile
125 120.18 or 129.81 0. 0385 5th
500 378.5 or 621 0.243 50th
900 394.2 or 1405 0.562 95th
Background: Trees, snow and asphalt pavement N = 244 Table 16 Vehicle Contrasts from Luminances Measured in the Field
Vehicle Contrasts during Daylight Conditions Vehicle Contrasts during Dusk Conditions
DAYLIGHT (4:00 P.M.) DUSK (8:30 P. M.)
Background 1 Vehicle Vehicle Contrast Background Vehicle Vehicle Contrast Luminance Color of Car Luminance Luminance Color of Car Luminance Bt - Bt „ Bt -*b (foot Lamberts) (Subjective) (foot Lamberts) c » —...... - (foot Lamberts) (Subjective) [foot Lamberts) C ------Bb Bt Bb Bb Bt Bb
1850 Blue 1400 -0.243 White 28.0 0.167 1850 White 4000 1.162 24.0 Purple 7.5 -0.750 1850 Light Blue 600 -0.676 30.0 10.5 -0.700 1850 Half White 550 -0.703 35.0 Azure Blue -0.821 1650 Black 660 -0.600 42.0 Azure Blue with 7.5 1650 Maroon 1300 -0.212 White Top 1650 Blue 1200 -0.273 42.0 Gray 25.5 -0.393 1650 Half White 2200 .333 42.0 Red/Orange 18.5 -0.560 1650 Blue 1700 .030 30.0 Dark Blue 15.0 -0.500 1650 Tan 1000 .152 22.0 Rod 20.0 -0.091 1700 Blue 1600 -0.059 19.5 Light Blue 17.5 -0.103 1700 Blue 1350 -0.206 15.5 Oil Green 12.0 -0.226 1700 Tan 2050 0.206 18.0 Light Yellow 12.0 -0.333 1600 Red 2600 0.625 12.5 Dark Blue 7.5 -0.400 1600 Green 1600 0.000 10.0 Steel Gray 5.0 -0.500 1600 Rod 1150 -0.281 10.0 Red-Purple 5.0 -0.500 -0.474 1650 Maroon 1100 -0.333 9.5 Rod 5.0 1650 Orange 2000 0.212 8.0 Brown 3.6 -0.550 -0.333 1500 Blue 2000 0.333 9.6 Red 6.4 1500 Purple 1000 -0.333 1500 Green 1200 -0.200 Mean- 0.435 SD - 0.213 1650 Yellow 3100 0.879 Min - 0.091 Max - 0.821 Mean> 0.373 SD -0.295 Min -0.0 Max -1.162 NOTE: The statistics glren are based an absolute contrasts.
NOTE* The statistics given are based on absolute contrasts. coto 99
7 v
3.75" (9.52 cm)
6 / 10' Each digit is a seven segmented 7 5" Light Emitting Diode (LED) of <19. 05 cm) 6/10" character height.
Figure 26: Display Dimensions.
The digits were electrically wired so as to produce an ’8' or a ’6* when lighted. Presentation of digits was controlled by the PDP-8/L computer so that any one of the eight configurations (all 8's, all 6's, any one of the six posi tions as a '6' and the rest as '8') could be presented with a desired display on and off time.
T ask:
With the above display, the following types of tasks can be generated:
• Tasks of cognitive nature in which the subject may be required to do arithmetical operations on the digits of the display.
• Tasks involving memory in which the subject may be required to respond to the occurrence of a particular sequence of dis play presentations.
• Tasks involving the determination of threshold luminance of the digits for their recognition.
• Tasks involving visual search in which the subject may be re quired to search for a dissimilarity in the display and make an appropriate response.
Since the visual information acquisition in driving involves more of visual search than other types of task nature, the foveal tasks were designed to in volve visual search only. This was achieved by classifying all the eight possible stimuli (digit combinations) as dissimilar and similar in the following way: 100
Possible No. Task Type Display similar if Display dissimilar if of Stimuli
1 all digits are 8's any one of the six 8 o r digits appear as "6" all digits are 6's and the rest as "8"
2 all the digits show any one of the six 7 an *8' digits appear as '6' and the rest as '8'
In both types of search task the subject's response was in a binary mode. If the accuracy of the subject's response is the only performance criterion, the subject may be required to make oral responses,,
If latency of response is also considered as a criteria for judging the performance, then the subject may be given a set of response buttons which can generate electric pulses for recording the elapsed time during a response. In any case the response mechanism was considered to be simple enough to pre vent its interference with the subject's visual search for information.
Modeling the Task Demands
Task Demands:
The task demands can be modeled under the information theory paradigm.
The stimulus uncertainty H(s) can be given by . _ pj log2 (l/Pj) where N 1 “ 1 is the number of possible different display (stimulus) conditions and Pj is the probability of occurrence of stimulus condition i. The stimulus uncertainty can be varied by any one of the following methods:
1. Varying the Number (N) of Possible Stimuli: By increasing N,in general, stimulus uncertainty can be increased. With the display device available the maximum value of N is limited to 8.
Let three levels of stimulus uncertainty be produced by using 6, 4 or 2 dis similar stimulus presentations. The stimulus uncertainty associated with each level may be calculated as shown in Table 17. It can be seen that the Table 17
Stimulus Uncertainty Associated with Different Number of Alternatives in the Foveal Task
Level of Pi, Probability of Occurrence of Different Stimuli Number Hjs) Task Assumption Tt.sk of Similar Dissimilar f Pt'lo SS^ Type about Pi Demand All All A '5' in P osition Stimuli 8's 6'8 I 2 3 4 5 6 N t'1 bit, v i
1 Occurrence 1 1/4 1/4 1/4 1/4 • — 4 2.0 of similar & dissimilar 2 2/8 2/8 1/8 1/8 1/8 1/8 -- 6 2.5 display are equiprobable 3 3/12 3/12 1/12 1/12 1/12 1/12 1/12 1/12 8 2.7925
2 (« 1 2/4 - 1/4 1/4 - - - - 3 1.5
2 4/8 - 1/8 1/8 1/8 1/8 -- 5 2.0
3 6/12 - 1/12 1/12 1/12 1/12 1/12 1/12 7 2.2925
3 All stimuli 1 1/4 1/4 1/4 1/4 - -- _ 4 2.0 are equi probable 2 1/6 1/6 1/6 1/6 1/6. 1/6 - - 6 2.5
3 1/8 1/8 1/8 1/8 1/8 1/8 1/8 1/8 8 3.0
4 it 1 1/3 - 1/3 1/3 - - - - 3 1.585
2 1/5 - 1/5 1/5 1/5 1/5 - 5 2.3219
3 1/7 - 1/7 1/7 1/7 1/7 1/7 1/7 7 2.8074 102 three levels of stimulus uncertainty obtained by varying the number of stimulus alternatives can be associated with a change in the display size. Levels 1, 2, and 3 would require a 2,4 and 6 digit display which can be ar ranged as shown in Figure 27.
1^—0 .24°-H £------3°------*1
The visual angles are with respect to display distance of 6 feet from the subject.
Figure 27: Three Displays Producing Different Area for Search
If the luminance of each digit should be adjusted to the threshold value for discrimination by fovea only, then the subject has to increase his search activity as the display size increases. Thus variation in the number of possible stimuli produces a simultaneous change in the display search area.
2. The second method of varying the stimulus uncertainty is by varying the probability of occurrence of each stimulus. An example of such changes in the stimulus uncertainty can be seen in Table 17 under different assump tions for pj.
3. The third method of varying the task demands is to consider the stimulus uncertainty rate. The stimulus uncertainty rate is given by the amount of uncertainty per stimulus presentation divided by the time per presentation. Therefore, by controlling the display "on time", the desired uncertainty rate can be achieved.
Subjects Response Console
The subject's response console (see Figure 28) consisted of: < 1) a steering wheel mock up placed in front of the front screen.
2) a seat which could be adjusted to the test subject's height. 103
mm projector \ with 0.5" lens
Eye position
■Chin rest Steering wheel with push seat buttons 44"
72"
1 2 4 "
NOTE; The arrows Indicate the direction of adjustment for the chin rest and the seat to maintain constant viewing distance from a constant eye level.
Figure 28: Schematic Diagram of the Subject’s Console, Front Screen and the Projector 3) a chin rest with telescopic arrangement for adjusting its height according to the subject’s height.
4) four response buttons fixed to the steering wheel (two on the left and two on the right).
5) a head phone connected to a tape recorder/player capable of playing highway driving noise. and 6) biometrics eye movement monitor for recording the hori zontal eye movement of the subject.
The subject’s chin rest was positioned such that its distance from the center of the two screens was always 72 inches.
The two response buttons on the right side of the steering column were used for responding to the condition of the display, i. e. the subject pressed the right top button if all the digits in the display showed 8’s. Otherwise, the subject pressed the right bottom button. The two response buttons on the left were used for detection and discrimination of peripheral targets. The purpose of playing the highway driving noise through a set of head phones was to mask the noise produced by the projectors.
Experimenter’s Console
The experimenter's console consisted of the following controls:
•power on/off switches for the film projectors and slide projectors
• the variacs for the control of the wattages of the projection lamps
• a light dimmer for the control of general lighting of the room
•power on/off switches for the foveal display, turntable, micro switches on the turntable, response buttons on the steering wheel, eye movement monitor, and the oscillograph chart re corder
• a trigger switch to initiate the program in the computer
• remote control for the slide change and focusing 105 Computer-Laboratory Set Up Interfacing
The objectives of interfacing the laboratory set up with the computer are:
1. for the control of the foveal display to produce desired task demands for the test subject. and 2. to record data on line regarding the foveal task conditions, the subject's response to foveal task and peripheral task, and movement of the turntable on a common time base.
The elements of the laboratory set up that were interfaced with the PDP- 8L computer and the type of data recorded are shown in Figure 29c The com puter software utilized the built-in clock to record all the time data. While the data of the foveal task performance is straight forward in term s of accuracy and latency, a brief description for the computation of peripheral response angle is given below: In manipulating peripheral task data, it was assumed that the turn table speed was constant for a given trial while moving through the given test region angle. Then let the time interval between the signals from the left and right switches of the turntable be tT seconds. During this time let the target image on the screen move from to a 2 degrees. Let the subject's response to peripheral target occur after tp seconds from the instant the image moves in ward from a j. Then the peripheral response angle a p corresponding to the subject's response can be determined as follows: (See Figure 30.)
Slide projector movement time within the test region = tT secs.
Test region = a ^ - ag degrees
Target angular velocity = (a ^ - a 2) degrees/sec.
Subject's response time to peripheral target 1 when it was at ap J tp secs*
Target movement during tp secs. = (a^ - (I 2 ) * tp tT Peripheral response angle COMPUTER ELEMENTS OF THE SIGNAL (Signal T^pe and LABORATORY PURPOSE Flow Direction) SOFTWARE INPUT To Control Display On/Off Time A <—D Programs for: Foveal Display and the A D 1) Reaction Time C Display Sequence A D 2) Foveal Task Only A D 3) Peripheral Task Only and Response Buttons To Record the Response Latency A ~ ) D 4) Foveal and Peripheral (Foveal Task) and Accuracy A—D PDP-8/L Task, Data for Display C Control, Number of Trials DATA OUTPUT Computer
Response Buttons To Record the Time of Detection A •— p D Foveal Task (Peripheral Task) and Discrimination Response A -» D with C 1) Display Task Conditions Disk 2) Trial Sequence 3) Response Accuracy Turntable (Left and To Record the Time for a Given A -» D Storage 4) Response Latency Right Switches) Angular Movement of the Slide A—> D Peripheral Task C Projector 1) Turntable Movement Time 2) Response Latency and Accuracy Program Trigger To Initiate or Terminate Data A->D C Switch > Collection
Figure 29: Schematic Diagram Showing the Interface between Laboratory Elements and the Computer 107 10°
Front Screen Left Screen / T arget E ndsv Inward \ Movement tim e = t-p
^ Subject Responds to Target Subject tim e = t
Test Region = (a^ - a 2)
al Target Begins Inward Motion tim e = 0
[ ■ ( a j- a g ) • tp I Peripheral ResponseAngle a p = a1-L tT J
degrees from straight ahead position
Figure 30: Schematic Diagram Showing the Visual Field Determination CHAPTER 5
RESEARCH STUDIES: PURPOSE, PROTOCOL AND RESULTS
This chapter describes the various experiments conducted in this research under the following three sections:
1. Subject Baseline Data Studies
2. Pilot Experiments on Peripheral Detection Angle (PDA)
3. Major Experiments on Peripheral Detection Angle (PDA)
In the baseline data studies, efforts were made to first screen the subjects for vision abnormalities and record their normal visual field. In the laboratory, each subject was tested to determine the following:
1. Simple reaction time, thumb travel time, and response time to a light signal with no uncertainty using the steering wheel mock up and response buttons.
2. A critical display "on time". This time is defined as the shortest "on time" for the digital display in the foveal task under which the subject can perform with no more than two errors in the foveal task without peripheral detection task.
3. Threshold target visibility. This is defined as the average target contrast required for detection of a stationary highway target exposed with its leading edge at 90° eccentricity.
In the pilot experiments on the Peripheral Detection Angle (PDA), information on the following aspects were sought in order to aid the design of an experiment to study the effect of levels of foveal loading and target visibility on human peripheral detection angle (PDA).
1. The variation in the peripheral detection angle (PDA) with repeated observations.
109 ♦ 110 2. The effect of testing in the laboratory on two different days on the peripheral detection angle (PDA) data.
3. The effects of no foveal load and foveal load on the peripheral detection angle (PDA) under one target visibility conditions.
4. Guidelines for the selection of the levels of foveal task demands and target visibility levels.
5. The effect of different target images contained in different slides.
To obtain experimental evidence to the above issues a pilot experiment was conducted in four stages. The description of each of the stage and the results are given in the section entitled "Pilot Experiments".
In the "Major Experiments" on peripheral detection angle, the effects of three levels of foveal load and three levels of target visibility on subjects' peripheral detection angle (PDA) were investigated. The levels of foveal loads and target visibility levels were based on individual subject's critical "on time" for the foveal task and threshold for target visibility. | 1 Figure 31 provides an overview of all the phases of the experimental efforts of this research.
Eye Examination of Test Subjects
P urpose:
The primary purpose of this part of the research was to assess the visual characteristics of ail the test subjects with the idea of:
1) selecting only those subjects who were free from any abnormalities in their vision.
2) to obtain data on the subject’s normal visual field using perimeters and two different target sizes. The data on subjects' normal visual field will be used later to correlate with the subject's peripheral detection angle (PDA) data obtained in the laboratory.
Experiment:
< The subjects were tested at the Optometry Eye Clinic of the University. The guidelines for testing and subject selection were as outlined below. Studies on Baseline Data Pilot Experiments Major Experiments on PDA \ ItlE LlM IN A ItY SELECTION OF SUBJECTS 1 S T A C E X : 1 Data: A c*. Sex. Height, D riving Experience E ffect o f repeated observations and testing ea different PDA EXPERIMENT 1 days on PDA 2 young subjects, three foveal loads, four visib ility levels 1 subject. 2 days. 30 observations*d o foveal load* one 1 target visibility EYE EXAMINATION OF SUBJECTS
Subject screening, Norm al visual Hold 1 r S T A C E n : Effect of Foveal Task Loading on PDA PDA EXPERIMENT 2 i 1 subject. 2 days. 39 observations per day. one target 3 young and 3 aged subjects, REACTION* TIM E EXPERIMENT v is ib ility , no foveal load and high load three foveal loads, three visib ility levels
Dux: Simple reaction tim e. Travel tim e. Response tim e
S T A C E m : • E ffect of three levels o f foveal load and three levels of < f visib ility on PDA FOVEAL TASK PERFORMANCE SPECIAL ANALYSIS 1 s u b j e c t . 1 t a r g e t C ritica l "on time** fo r subject loading Intercorrolatlon of baseline data w ith PDA data. I Interpretation o f tho overall results and conclusions* S T A C E I V : 1 Effect o f four different target slides at two varies THRESHOLD OF VISIBILITY settings on visual flold
Data on thrrahold o l visib ility (or dctoettoa 1 subject, two fovoal loads, two replications o( .latlonar? target at 90°
V
Figure 31.—Schematic Diagram Showing the Overview of all the Experimental Phases of this Research. Assessment of the Visual Characteristics of the Subjects:
Each subject must undergo the sequence of tests and measurements outlined below:
1) Tests for pathology of the eyes including, among other things, test for glaucoma (older subjects), visual fields, and the ophthalmoscope.
2) Careful check on the normal blind spot and other scotomatous gaps in the horizontal meridan of the visual field paying special attention to the edges of the spectacle lens when such lenses are to be used in the experiments. For plus lenses, the target disappears; for minus lenses, the target doubles.
3) Visual acuity (O.D. and O.S.)
4) Phoria test.
5) Refractive error of each eye.
6) V.A., O.D., andO.S. with new Rx.
7) Axial power of the old lenses.
8) Curve (or curves) on the ocular surface of each lens.
9) Distance from the ocular surface of each lens to the eye.
Disqualifying Characteristics:
1) Wearers of contact lenses are disqualified because of the difficulty in assessing the effect on peripheral vision.
2) Freedom from pathology which effects peripheral vision.
3) Each subject must have 20/25 V. A. or better in each eye with the glasses to be worn in the experiment.
4) The phoria must not exceed of right or left hyperphoria and must not exceed 3 of exo or esophoria.
5) Squinters and one-eyed people are disqualified. 113 To obtain the normal visual field of subjects, TOPCON Perimeter was used. The details of targets used are as follows:
•Larger target: size - 2.5 mm^ (disk diameter 1.78 mm) luminance - 100 abs.
•Smaller target: size - 0.25 mm^ (disk diameter 0.56 mm) luminance - 100 abs.
•Background luminance: 31.5 abs.
R esu lts:
The field data for the horizontal meridian for each subject is given in Table 18.
Simple Reaction Time
P urpose:
The purpose of this experiment was to determine the simple reaction time, thumb travel time, and the total response time to a light signal with no associated stimulus uncertainty. This information regarding each subject was felt necessary for the following reasons:
• The test subjects used a special console (steering wheel mock up) with response buttons in different phases of this research. Know ing the basic response time (response time with no stimulus un certainty), it is possible to adjust any other response time which contains an element of search time. For example, Bloomfield (1973) used the following methodology to determine the search time:
Search time Total response time Total response time under uncertainty = under uncertainty - under zero uncer- H (s) H (s) tainty
Experiment:
The subjects were tested using the digital display and the response buttons of the simulation set up. The front and the side highway scenes were projected during testing. The subjects were asked to press a response button whenever the lights of the display in front of them were on. 114 Table 18
Normal Visual Field Data in the Horizontal Meridan in Visual Degrees
Smaller Target Larger Target Subject Left Eye Right Eye Left Eye Right Eye T N N T T N N T
SOI 63 32 26 65 76 53 54 73
S02 48 27 50 63 90 58 62 88
SOS 61 31 35 54 68 52 58 68
S04 70 52 58 68 90 -- 90
SOS 56 35 42 42 70 50 53 71 '
S06 48 36 38 32 68 58 56 63
S07 39 28 27 42 90 - - 90
S08 64 38 40 58 65 48 40 71
S09 30 28 28 40 90 55 64 88
S10 60 43 43 53 90 - - 90
Note: Normal visual field refers to the peripheral detection sngle corresponding to the detection of a small birght disk with the eye fixated at a central point.
See text for details on target size, luminance.
« 115 The subjects used two push buttons during the experiment. One of them was kept pressed by the subject to initiate the signals. The second was used to respond. Hence, the task was subject paced.
Each subject was given sufficient training before data collection. Thirty test trials were given to each subject. Figure 32 explains schematically the experimental protocol and the computer-iaboratory set up interfacing for data collection.
Subjects:
Ten subjects (6 young and 4 aged) were tested to collect data on their reaction time, thumb travel time, and total response time.
R esu lts:
The mean and standard deviation for the reaction, travel and response times are given in Table 19 by subject, age group, and overall. The data was analyzed using the SAS REGR (regression) procedure with age group and subjects within the group as independent variables. The analysis indicated that age group was a significant factor affecting reaction time and response time = 0.0001). It can be seen from Table 19 that young subjects responded more quickly than aged subjects. The analysis also indicated that the thumb travel time for the groups was not as significant (p = 0.109) as the reaction time. The Analysis of Variance (ANOVA) tables from the regression analysis for reaction, travel and response times are given in Tables 32, 33 and 34 in Appendix A.
Foveal Task Performance
Purpose:
The primary objective of this experiment was to determine for each subject a high level of loading in the foveal task at which the subject can perform with a minimum of errors in the foveal task.
Experiment:
The foveal task was provided using the lighted digital display described in the chapter entitled "Laboratory Set Up". The display with 6 digits provided a 6° x 3° area for the subject to search foveally during the task. It is to be noted that in driving, drivers’ eye fixations normally lie in this region around the focus of expansion region. The display provided two different stimulus conditions with equiprobability: in one, all the six lights showed 8’s, and in the other, one of the six lights showed a "6"and the rest 8's. The subject's 116 ■Initialization of Computer Program
Instructions to Subject
Start of the Program (switch on the experimenter’s console) 10 secs. T ± Audio Signal to Alert the Subject Computer to Wait Subject Starts the Trial Sequence (Response T button 1 kept pressed) 2 secs. • Display on (all 8's lighted) Reaction 4 T Time Response Subject Reacts (button 1 released) Time T ravel 4 Time i J l Subject Responds (button 2 pressed; trial completed) Computer to Wait Subject Starts Sequence of Next Trials 2 secs.
- 4- Display on
/ Subject Responds (trial 30 completed)
•Audio Signal to Mark End of Data Collection
.Data Output on Teletype
Figure 32.—Schematic Diagram Showing the Computer Control for Data Collection in Reaction Time Experiment. Table 19
Data on Simple Reaction Time
Age Number of Simple Reaction Time Thumb Travel Time Total Response Time Subject Yrs. Trials MSECS MSECS MSECS Mean Std. Deviation Mean Std. Deviation Mean Std. Deviation
SOI 19 30 192.1 42.9 110.3 15.3 306.1 43.2
S02 22 30 306.8 25.6 129.4 13.91 436.2 27.5 S03 37 26 202.6 37.6 120.8 23.8 323.4 29.0
S04 23 30 223.5 29.5 104.1 8.6 327.6 31.6
S05 25 30 269.8 31.1 99.6 9.1 369.5 31.1
S06 29 30 213.5 46.2 119.0 53.6 332.5 30.7
S07 56 27 254.6 30.2 125.7 23.8 380.3 37.2
S08 57 30 287.3 58.8 145.8 16.0 432.8 63.7
S09 69 30 316.9 54.6 127.9 17.4 444.8 58.5
S10 65 30 241.1 38.7 74.0 9.6 315.1 40.5
SOI to S06 Young 176 235.4 54.4 113.7 27.5 349.8 54.3 S07 to SIO Aged 117 275.5 55.5 118.1 32.1 393.6 72.8
SOI to SIO Overall 293 251.4 58.2 115.4 29.4 367.3 65.8 118 task was to press the right top response button for the "all 8"condition and right bottom response button for the "one 6" condition. The loading in the task was achieved by decreasing the stimulus "on time". The laboratory set up was interfaced with the PDP-8/L computer to provide display control and data collection. Figure 33 explains schematically the data collecetion procedure for a given loading condition. The software used for data collection had one limitation, i.e ., it recorded only one response for each stimulus presentation. Hence, if the subject's response time was longer than one cycle time, (on time plus off time) then the response was recorded corresponding to the next stimulus cycle. This problem did not occur with longer cycle times. However, this software limitation was eliminated in the later usage of the program.
The subjects were given the following instructions:
"In this experiment, information regarding your speed and accuracy of response to a given signal will be gathered.
When you hear an audio signal, you are requested to sit straight with your chin resting on the chin-rest, head facing the six-digit display in front of you and your hands gripping the steering wheel. Position your right thumb conveniently so that you can press any one of the two right side response-buttons on the steering wheel.
After 10 seconds from the audio signal, the display will be in the on and off mode at regular intervals of time. When the display is lighted with all 8's, press the top button once as quickly as possible. When one of the lights shows a "6" and the rest as "8's", press the bottom button as quickly as possible.
Repeat the above performance until you hear an audio signal to mark the end of the trial. If you have any questions, the experimenter will provide clarifications."
During the experiments, moving highway scenes were projected on the two screens. The subjects were adapted to the experimental lighting conditions gradually. The subjects were given a series of test sessions with 20 trials in each session. The first session started with a fairly easy task loading, i.e ., high "on time" such as 2500 msecs. for aged subjects and 2000 msecs. for young subjects. After each session, the results obtained on the teletype of the computer were quickly reviewed for number of errors. If the number of errors was two or less, the "on time" was reduced by 250 msecs. in the next session. If the errors were more than three, then the subsequent sessions utilized increased "on time". The experiment was stopped around the "on time" 119 Initial Input to the Program
Starting the Experiment (Experimenter1 s Control) 10 Secs.
Audio Signal to Alert the Subject 10 Secs. T Begin Trial l Display Off Time 250 M secs. Display On (Display Condition According to Predetermined Random Sequence) Display "Oh. Tim e" Subject’s Response (Type) Response— Time -Display Off Off Time F (250 msecs. Display On (2nd T rial)
Z T Display Off (End of Last Trial) 10 Secs. Audio Signal (End of Test Session)
Data Output on the Teletype
Figure 33. —Schematic Diagram Showing the Computer Control and Data Collection in the Foveal Task Experiments 120 with which testing was first started. Hence, a typical schedule of test sessions involved "on times". For example, for young subjects, the "on times" were: 2000, 1750, 1500, 1250, 1000, 750, 500, 750, 1000, 1250, 1500, and 1750 msecs. The objective of repeating the "on times" during the later part of the testing was to verify the stability of subject's performance at a given condition. Each subject was tested approximately for 12 sessions within a period of one hour.
The data were recorded in the following format:
Trial Number: 1 to 20
Display Type: All 8's or one 6
Response Type: 1 for top button; 2 for bottom button
Response Time: Time in msecs. from the onset of the stimulus
R esu lts:
The data was analyzed to obtain the following statistics for each session:
1) For all the 20 trials:
a) Total Number of Errors (including wrong response and m isses)
b) Error Probability
c) Mean Response Time (considering only the right responses)
d) Standard Deviation of Response Times
2) For the 10 trials with "all 8's"
The four measures as in 'a'
3) For the trials with "one 6"
The four measures as in 'a'
A sample output of the data analysis program is shown in Table 20. «
The important result that will be used in the subsequent experiments is on the number of errors in a given session of 20 trials. The session with the Table 20
Sample Output from the Analysis of the Data on Foveal Task Performance
£QVFAl TASK PERFORMANCE
JOE STAFFORD, 6-2-76. PILOT EXPT. FOVEAL LOADING, 4 DIGIT DISPLAY
DISPLAY CONDITIONS PERFORMANCE ItOVFR-ALL) performance (ALL 8) PERFORMANCE (ONF 61 HO. nv OFF NO. NO, R-TIMF R-TIME NO. NO. P—TIME . R-TIM8 NO. VO. 9-T1MF R-TIMF OF TIM? TIM? OF CF moir MEAN S.D. OF OF Error MEAN S.D. OF ne E°°CR Mp AN 5.0. DIGITS MSECS MSECS TRIALSERRORS PR OB. MSECS MSECS TRIALS ERRORS PROS • MSFCS MSECSTRIAL?ERRORS PROS. MSFCS MSECS 4 2000 250 20 0 0.0 H25.50 123.43 1(1 0 0.0 897.50 114.84 10 0 0.0 86.35 4 1750 250 20 0 0.0 827.30 108.02 10 0 0.0 890.40 95.31 10 0 0.0 7F4,?0 96.33 4 1500 250 20 0 0.0 .... 807. 05 105.58 10 0 0.0 834.20 rr .so 10 0 0.0 770.00 17?.64 4 1250 250 20 0 0.0 753.00 147.56 10 0 0.0 830.00 142.e6 10 0 o.o 676.00 111.29. 4 _J000 250 ?0 0 0.0 692.35 124.44 10 0 0.0 781.20 135.46 10 0 0.0 . 603.50... 47-6? 4 750 250 20 0 0.0 607.60 77.37 10 0 0.0 664.40 60.50 10 0 o.c 560.30 25.24 4 500 250 ?o 0 0.0 559.00 100.17 10 0 0.0 5” .70 IDE.97 10 0 0.0 6P4.00 88.37 4 400 250 17 5 0.294 470.33 117.22 'Q 2 0.222 471.71 137.73 8 3 0.375 468.40 96.54 4 500 250 IP 0 0.0 490.58 158.53 0 0 0.0 533.56 109.56 10 0 0.0 469.00 193.30 4 750 250 20 0 0.0 654.45 91.85 10 0 0.0 701.90 85.05 10 0 0.0 607.00 74.67 4 1000 250 2.Q 0 0.0 740.35 92.10 10 0 0.0 . . R04.OO 63.no 10 0 0 .0 716.80 104.73 4 1250 25C 20 0 0.0 75*-.70 97.44 10 0 0.0 830.10 69.47 10 0 0.0 679.30 62.23
4 1500 250 20 0 D jlO 802.90..147.0? 10 0 0.0 003.30 136.05 10 0 0.0 704.*0 71.86 4 1?53 250 2C 0 0.0 847.10 115.83 10 0 0.0 902.10 59.81 10 0 0.0 792.10 124.26
4 2000 250 20 0 0..0 360.90 135.45 10 0 0.0 948.60 07.41 10 0 0.0 . 777.70 114.46 121 shortest "on time" in which the number of errors committed was no more than two indicated in Table 21 for each subject. This "on time" can be called as the critical "on time" for each subject. At this level of loading, it is reasonable to assume that the subject’s spare capacity for information processing is kept to a minimum level. It can be observed from Table 21 that young subjects had critical "on time" of 1000 msecs. or less and aged subjects had critical "on tim e" of 1250 m secs. or m ore.
It was also observed that the response times under the "all 8" condition were longer than the response times under the "one 6" condition. This result indicates that in the "all 8" condition, every digit has to be verified as "8" and in the "one 6" condition it is possible, even before searching all the digits, that a 6 might be encountered and the search stopped. Individual subject's data on mean response time, standard deviation of response time and the total number of errors out of 20 trials are given in Figures 56-64 in Appendix A.
To investigate the effect of display size on the response time, data was collected using subject SOI and the size of the display as a 2-digit and 4-digit display. The total stimulus uncertainty for the two and four digit displays were 1.5 and 2.0 bits. The results are shown in Figure 34. It can be seen that the mean response time and the standard deviation of the response times were smaller for the two digit display when compared with the four digit display.
Table 21
Foveal Task Performance Critical "On Time" by Subjects
Critical Time-Msecs. (2 or less errors Subject out of 20 tria ls) rsoi 750 S02 1000 S03 - Young *< S04 750 S05 1000 S06 1000
rS07 1750 S08 1500 * S09 1750 hSlO 1250 Time in Seconds 0.9 - 0.9 1.1 1.2 0.7 0.8 1 .6" 6 0. 0.1 0.5 0.0 0.2 0.3 - .4 0 . 0 - Figure 34. —Mean and Standard Deviation of Response Times as a Function of of Function a as Times Response of Deviation Standard and —Mean 34. Figure . 17 1.5 1.75 2.0 Stimulus on Time for Four and Two Digit Displays for Subject SOI. Subject for Displays Digit Two and Four for Time on Stimulus Stimulus On Time - Seconds - OnStimulus Time Tour Digit Display (2.0 Bits) (2.0 DigitDisplay Tour Test Conditions Test Subject - SOI - Subject Seconds 0.25 OffStimulus- Time Two Digit Display (1.5 Bits) (1.5 TwoDigitDisplay
€* <€>* 1 - -O— -©- —© -©- -O— -
.5 2.0 1.75 Standard Deviation Standard of Times of Response Response
Time Mean 123 124 Threshold Visibility
Purpose:
The general purpose of this experiment was to determine the subject's threshold for target visibility when a stationary target is presented in the periphery. Target visibility refers to the target contrast determined by using the average luminances of the target and the background. The specific objectives of the experiment were as follows:
1) To determine the target visibility threshold for each subject when the test target was exposed at 90° eccentricity.
2) To determine the effect of full and partial target exposure on visibility threshold, at 60° angular location.
3) To determine the effect of two different angular locations (90° and 60°) on visibility threshold.
Experiment:
This experiment used the simulation set up described in detail in Chapter 4. Subjects were seated facing the front scene with their chin resting on a chin-rest. Subjects' eye level was adjusted to be 44" above floor level. The luminance of the front and left side highway scenes was always kept at 0.155 ft. Lamberts. The vehicle image was superpositioned for the left scene at angular locations of 90 and 60 . The average luminance of the vehicle image was controlled by controlling the lamp wattage of the slide projector.
Three different conditions of target presentation were used for the threshold study. See Table 22 for details of the three vehicle images.
During the testing, the subjects were required to fixate their eyes at .the center of the road ahead of them. The experimenter increased the luminance of the vehicle image gradually and slowly. The subjects made an oral response when they first detected the vehicle in their periphery. Immedi ately the reading of the variac setting was recorded and the image luminance was reduced to zero. The above procedure was repeated 10 times for each of the target conditions described in Table 22. After each 10 trials, the subject was given 3 minutes of rest period. Before beginning the testing in each test period, the lamps of the scene projector were switched on and then the general room illumination was gradually decreased to adapt the subjects to the scene luminance without eye strain. At the end of the trial, the room lighting was gradually increased before cutting down the scene luminance. 125
Table 22
Vehicle Image Size and Visual Angle Subtended at the Subject's Eye
T arget Visual Angle Presentation Image Size Subtended (Degrees) Condition Leading Edge Vertical from Horizontal from and Portion Floor Level Outer Edge of V ertical Horizontal of Vehicle Left Screen
Leading Edge at 90° and 12” to 36” 20” 16° 12° Front of Car
Leading Edge Front 18” to Front - 45” to F ro n t-16° 12° at 60° and 40" - Body 60” Body - 60" Body-24° 42° Full Exposure 18" to 50"
Leading Edge o CO at 60° and 18” to 40" 45” to 60" iH 12° Front Only
Eye Height = 44" above floor level
The target was the image of a Volkswagon. 126 Ten subjects (six young and four aged) were tested in this experiment. The whole experiment lasted 15 minutes. Subjects were also given a few trials before data collection.
R esu lts:
The results on the threshold visibility for all the ten subjects under three target conditions are shown in Table 23. The data was analyzed using SAS REGR procedure with the following as independent variables:
•Age
•Subjects within Age
•Cl representing the contrast between the means of data under partial exposure at 90° and 60°
• C2 representing the contrast between the means of data at 60° with full and partial exposure
The ANOVA from the regression analysis is given in Table 35 in the Appendix A. From the Table it can be seen that:
1) at 60° eccentricity, the average threshold visibility for partial target exposure is slightly higher than for full target exposure, (p <0.001)
2) for partial target exposure, the threshold visibility at 90° is significantly greater when compared with the threshold visibility at 60°. (p <0.001)
3) there is a general tendency for aged subjects to exhibit a slightly higher threshold visibility at all conditions. (p <0.001)
The cumulative percent of trials in which detection was made as a function of target visibility for the three target conditions are shown in Figure 35• From the Figure it can be clearly seen that the threshold for target visibility at eccentricity 90° is distinctly higher than at 60° eccentricity.
Also one can see from the Table that the standard deviation of the threshold values was higher at 90° eccentricity than at 60° eccentricity. Table 23 127 Threshold Visibility Data for Three Different Conditions of Target Presentation in the Laboratory for Ten Subjects Vehicle Leading Vehicle Leading Vehicle Leading Edge at 90°-Front Edge at 60°-Full Edge at G0°-Front of Car Exposed Vehicle Exposed of Car Exposed Subject Age N Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.
SOI 19 10 1.458 0. 361 0. 327 0. 073 0.356 0. 052
S02 22 10 1.755 0.145 0.418 0.142 0.458 0.139
S03 37 10 2. 077 0.521 0. 374 0. 087 0.426 0.125
S04 23 10 0.871 0.251 0.223 0. 072 0.327 0.116
S05 25 10 1.138 0.308 0.246 0. 041 0.233 0. 032
S06 29 10 1.448 0.251 0.311 0. 063 0.311 0. 052
S07 56 10 1.226 0.177 0.491 0.132 0.436 0.098
S08 57 10 1.542 0.228 0.503 0. 072 0.560 0.136
sod 69 10 1.987 0.276 0.494 0. 082 0.560 0.150
SIO 65 10 1.103 0.222 0. 355 0. 084 0. 377 0. 084
SOI to Young 60 1.458 0.504 0. 316 0.107 0.352 S06 0.118
S07 to SIO Aged 40 1.464 0.410 0.461 0.111 0.483 0.140
SOI to All 100 1.461 0.466 0. 374 0.129 SIO 0.404 0.143
Background Luminance = 0.155 ft. Lamberts
The target was the image of a Volkswagon.
Target Luminance - Background Luminance Visibility - Background Luminance
In the Simulation Set up
Target Luminance = Slide Luminance + Film Luminance
Therefore _____ Slide Luminance ______Visibility Background Scene Luminance Cumulative Percent of Trials In which Detection was Made. I 30 40 50 10 60 SO 90 0.0 iue 5 Cmltv Dsrbto o h Trsod Visibility Threshold the of Distribution —Cumulative 35. Figure 0.5 Data for Ten Subjects Ten for Data 1.0 re Visibility arget T 1.5 ) edn Eg a 60 at Edge Leading 2) ) edn Eg a 60 at Edge Leading 3) 1) Leading Edge at 90 90 at Edge Leading 1) ul Exposure Full ril Exposure artial P ril Exposure artial P .5 2 3.0 128 Pilot Experiment - Stage I 129
Purpose:
The objective of this experiment was to investigate the variation in the peripheral detection angle (PDA) with repeated observations over two days.
Design:
In this experiment one young subject was tested on two different days. In each day 30 observations were made on the PDA in the laboratory under no foveal load. The target was a bright Volkswagon (average luminance - 1.15 foot-Lamberts, the leading edge was located at 70" from the outer edge of the left screen covering a horizontal field of 50°, the vertical size of the image at the body was 26" starting at 14" above the ground level; the body subtended an angle of 20° at the subject's eye; the front of the car subtended an angle of 12.5° at the subject's eye). The highway scene produced about 0.155 foot- Lamberts luminance conditions.
P rocedure:
The subject was required to sit straight with his chin resting on a chin- rest and facing the center of the front screen. The highway scenes were projected as the subject was gradually adapted to the scene brightness. Instructions were given to the subject to fixate his eyes at the center of the roadway straight ahead and respond through a button the detection of a passing highway target on the left adjacent lane. The set up was interfaced with the PDP-8/L computer for data collection purposes. Each trial lasted for less than a minute. After five trials, three minures rest was given to the subject.- Within every five trials one catch trial with no peripheral target was introduced to check the false alarm response from the subject. Thirty observations were made on each of the two days testing.
R esu lts:
The procedure for the determination of PDA corresponding to the subject's response described in the laboratory set up chapter was adopted in a computer program to obtain the PDA data. Table 24 gives the results of Mean PDA for subject SQ2 in degrees from the eye position and the speed of the target. The speed of the target averaged to be 3.19 degrees per second for both days.. This speed, for a target moving in the left adjacent lane, will represent a relative velocity of 0.75 mph. Since the speed varied slightly between trials (std. dev. = 0.76 degree per second) a regression model was used to determine the significance of the dependent variables with speed as a covariable. REGR procedure of the SAS package was used with the following model statement: Table 24
Peripheral Detection Angle (Stages I & n, Pilot Experiment) PDA in Degrees Speed in °/ seconds Subject Day Foveal Load N Mean S.D. Low High c.v. Mean S.D. Low High C.V.
SOI 1 No 30 95.93 2.64 88.87 98.81 2.75 2.78 0.78 0.86 3.53 28.11
SOI 2 2 b its/sec 30 96.96 2.40 87.89 100.28 2.49 3.53 0.63 1.85 5.17 17.79 « SOI 1&2 No 60 96.25 2.52 87.89 100.28 2.62 3.16 0.80 0.86 5.17 25.29
S02 1 No 30 90.69 3.94 84.03 100.13 4.34 3.35 0.83 1.74 5.75 24.91
S02 2 No 30 90.41 2.30 85.71 94.05 2.54 3.04 0.65 1.17 3.56 21.52
S02 1&2 No 60 90.55 3.20 84.03 100.13 3.53 3.19 0.76 1.17 5.75 23.78
o03 131 MODEL PDA = SPEED TRIAL DAY
The results showed that the PDA data did not vary significantly between days and also within a given day. Hence, this result suggests that the data from different days can be combined and analyzed together. The regression analysis also indicated that speed within the given range did not affect the PDA data. From Table 24 it can be noted that the PDA under no foveal load and average target visibility (contrast) of 7.4 averaged to 90.55 degrees with a standard deviation of 3.2 degrees. Figure 36 shows the plots of the cumulative percent of trials in which the target vehicle was detected as a function of the angular position of the target for days 1 and 2 separately. Inspection of the plot will indicate that the variability in PDA data on the second day was less than the first day. The ANOVA Table from the regression analysis and the residual plot are given in Table 36 and Figure 65 in Appendix A.
Pilot Experiment - Stage II
Purpose:
The objective of this experiment was to determine the effect of critical foveal load on the PDA under a given target visibility condition. The critical foveal load will refer to the loading produced by the critical "on time".
Design:
In this experiment one young subject (SOI) was tested on two different days. On the first day 30 observations were made under no foveal load. On the second day 30 observations were made under critical foveal load. For subject SOI the critical foveal load (from Table 21) was 750 msecs. "on time" and 250 msecs. "off time" with a six digit display. The display produced a total stimulus uncertainty equal to 2.0 bits. The target, target visibility and target size were the same as described in P’ilot Experiment - Stage I.
P rocedure:
The procedure of experimentation was the same as described in Stage I. The only difference was that the subject performed the foveal task on the second day. On the second day prior to the beginning of the peripheral task the subject was given sufficient training in the foveal task. The subject was instructed to respond to the foveal task by pressing right top button for the "all 8" condition and the right bottom button for the "one 6" condition. The subject was also instructed that their primary task was to produce error free performance in the foveal task. Cumulative Percent of Trials in which Detection was Made. ICO 20 30 40 70 SO 90 10 re nua Psto i ere fo tagtAed Position Ahead Straight from Degrees in Position Angular arget T iue3.—uuaie itiuino D Dt fr w Days Two for Data PDA of Distribution —Cumulative 36. Figure a 1 Day Sae I) (Stage 85 Day 2 Day Day 2 Day 9570 75 a 1 Day 10580 132 133 R esults:
Table 24 gives the results in terms of mean, standard deviation of PDA for no load and load conditions. The data was analyzed using a regression model with load and trial as dependent variables.
The ANOVA Table from the regression analysis and the residual plot are given in Table 37 and Figure 66 of the Appendix A. The results showed that there was no significant change in the PDA data due to increased foveal load. The target visibility for the experiment was about 7.4. Since this value of visibility used in the experiment is considered to be a high level of contrast it was concluded that under high target visibility levels such as 7.4 PDA is not affected by critical foveal load. Figure 37 shows the plots of cumulative percent of trials in which the target vehicle was detected as a function of the angular position of the target for the no load and load conditions. It can be seen from the plot that PDA showed a slight increase under load conditions. The increase is of the order of one degree on the average. This might be due to learning. The average target speed was 3.16 degrees per second which is comparable to 0.75 mph. relative velocity in the real highway environment when a car moves on the left adjacent lane.
Pilot Experiment - Stage III
Purpose:
The objective of this experiment was to investigate the effects of three levels of foveal load and three levels of visibility of a given target on the peripheral detection angle (PDA) of a subject.
Design:
In this experiment one subject (S03) was tested for three sessions. The design of the experiment is shown in Figure 38 • In each session observa tions using one level of foveal load were made. These observations correspond to the three levels of target visibility and five repetitions.
The foveal loading was provided by the digital display of the laboratory set up. Only two digits were used to provide two types of stimulus conditions. These were either "all 8" or "one 6" situations. The total stimulus uncertainty was 1.5 bits. The three levels of the foveal loads were as follows:
Level 1: No display Level 2: 1000 msecs. "on time"/250 msecs. "off time" Level 3: 750 msecs. "on time"/250 msecs. "off time" Cumulative Percent of Trials in which Target was Detected. 50 £0 30 70 80 90 10 0 Figure 37. —Cumulative D istribution of PDA Data for Subject SOI Subject for Data PDA of istribution D —Cumulative 37. Figure 70 re nua Psto i Dges rm tagtAed Position Ahead Straight from Degrees in Position Angular arget T Sae I udrN La n La Conditions. Load and Load No under II) (Stage 80 No Load Load No Dy 1) (Day 85 o Load No 95 Dy 2) (Day Load Load Load 10575 134 Target Visibility (medium) 0.7586 (high) 1.492 0.290 (low) iue3 —xei na Dsg o tg I Eprments Experim III Stage for Design ental •—Experim 38 Figure o od eim od ih Load High Load Medium Load No Sessions Repititions 135 136 For subject S03 with a two digit display it was found that 1000 msecs. "on time" provided critical foveal load.
The target used in this experiment was the same as described in Stage I experiments. But the three levels of target visibility chosen for this experi ment were much lower than that used in Stages I and II. This decision was made based on the evidence from Stage II that high foveal loads in the presence of higher target visibility conditions did not affect the PDA. Hence, in this experiment it was decided to investigate the effects of lower target visibility conditions on the subject's peripheral detection angle (PDA).
P rocedure;
The data collection procedure was the same as described in Stage n. The subjects were instructed to perform both the foveal task and peripheral task. In the foveal task they were to press the top button for the "all 8" condition and the bottom button for the "one 6" condition of the display. The subject's primary task was to produce error free performance on the foveal task. The peripheral task was to press one of the left side response buttons when the subject first detects a passing vehicle on the left adjacent lane. The display control and the data collection was done by the PDP-8/L computer. After every five trials the subject was given three minutes of rest. Catch trials were introduced at random to check the false alarm response of the subjects.
R esu lts:
The PDA data was analyzed for the main effects of target visibility and foveal load and for the visibility foveal load interaction effects using a two-way fixed factor ANOVA model (see Table 38 in Appendix A). The analysis showed that the foveal load and visibility of target significantly affected the PDA. Figure 39 shows the plots of the mean PDA as a function of foveal load for the three visibility levels. With the decrease in the visibility levels the PDA also decreased at all foveal loads (p_= 0.0001). Increase in the foveal load from zero to 1.2 bits per second did not produce significant decrease in the PDA. But an increase in the foveal load from 1.2 to 1.5 bits per second produced a significant (p = 0.002) decrease in the PDA. The standard error for all the observations was 2.89 degrees. Figure 40 shows the plot of the cumulative percent of trials in which the target was detected as a function of angular position of the target for the three target visibility conditions. Figure 41 shows the same type of plot for the three foveal load conditions.
i The most important result from this experiment was that low target visibility conditions produce a significant decrement in the PDA. The visibility 137
100
High Visibility a o OS Medium Visibility £ :>> W o8 90 u Low Visibility in ®
a>1 O £ 2 ft § © S 80
0.0 1.2 1.5
Foveal Load in Bits per Second
Foveal Load, Two Digit Display O n /O ff Time Seconds Low Medium High o.o/o.o 85.99(0.83) 87.09(1.89) 91.6 (1.43) 1 .0 /.2 5 84.84(0.44) 88.52(1.56) 90.74(2.18) .7 5 /.2 5 82.17(2.45) 87.15(1.80) 88.74(1.74) N = 5 for each average
Figure 39.—Mean PDA as a Function of Foveal Load at Three Different Target Visibility Levels. Cumulative Percent of Trials in which Target was Detected. 100 0 2 30 0 6 40 10 90 Figure 40. —Cumulative Distribution of PDA Data for Three Visibility Visibility Three for Data PDA of Distribution —Cumulative 40. Figure 0 70 re nua Psto i ere rm tagt ha Position Ahead Straight from Degrees in Position Angular arget T Levels (Stage m , Subject S03). Subject , m (Stage Levels 580 75 85 90 012 otLmberts foot-Lam 0.1529 = Luminance Background ) 1-492 3) agt iiiiy Levels Visibility Target ) 0.758 2) ) 0.290 1) 95 100 138 Cumulative Percent of Trials in which Target was Detected. 100 100 30 30 30 30 40 40 50 60 60 *70 10 SO 90 0
Figure 41. —Cumulative Distribution of PDA at Three Different Different Three at PDA of Distribution —Cumulative 41. Figure — nua Psto fTagt n ere fo Srih ha Position Ahead Straight from Degrees in arget T of Position Angular 70 1 1 1 1
75 oel od (ujc 0, tg III). Stage S03, (Subject Loads Foveal 1 V
80 1
2 3 85 1. 1 o Load No 1) 1y \\ 20 secs. m 250 \ \ \ Fva La _ Load Foveal 1 \ 1 oftme" tim "off \ \ ) 00 secs. m 1000 2) \ \ o i ” - e” tim "on | \ .. L ....
...... K . bits/sec b 1.2 K \ 2 \ l \ ) 5 msecs. m 750 3) \ Y \ S "ntme" tim "on VSX 20 secs. m 250 ^ \ \ "f i e" tim "off \ \ K K I ” I 1" 1 I T” “ I 90 ......
I I I 1 . 1 95
... . biss *“ c its/se b 1.5 I 1 100
105 139 140 levels used in the experiment were arbitrarily chosen but kept much lower than 7.4 that was used in Stage II experiment. It can be noted in the later experiments that the values for the visibility levels were based on the threshold visibility for each subject.
The data on PDA was transformed into a new value as PDAT = logjQ (110-PDA). This transformation was carried out to eliminate any skewness present in the data and test for the main and interaction effects. The ANOVA from the regression analysis and the residual plot for the variable PDAT are given in Table 39 and Figure 68 in Appendix A. It can be seen from the ANOVA Table that the results are unchanged.
The mean target speed during the experiment was 3.12 degrees per second with a standard deviation equal to 0.63 degrees per second.
Pilot Experiment - Stage IV:
Purpose:
The objective of this experiment was to investigate the effect of using different targets (slides) to obtain the vehicle image in the simulation set up upon the peripheral detection angle (PDA).
Design:
The experiment was a three factor experiment as shown by the design in Figure 42. The two foveal loads were provided with a two digit display and "on time" of 1500 msecs. and 750 msecs. The "off time" was kept constant at 250 msecs. The low and high target visibility were realized by arbitrarily setting the variac for the slide projector at 40 and 60 divisions. Because of the differences in the highlights and brightness among the four slides each slide produced different visibility conditions for the same setting of the variac.
For the four target slides used in the experiment the size of the image on the screen, luminance conditions and visibility levels produced are detailed in Table 25.
P rocedure:
The procedure for the data collection was the same as described in Stages II and III. The treatment conditions for the visibility-slide combination were exposed to the subject in one test session under one foveal load. The second session used the other foveal load. The complete experiment was replicated twice. One young subject (S03) was tested under the above protocol. 141
Low Foveal Load
High Foveal Load Slides for Targets
Low High Target Visibility
Figure 42.—Experimental Design for Stage IV Experiments
R esults:
The results were analyzed using a three factor fixed ANOVA model. The ANOVA Table is given in Table 40 in Appendix A. The analysis showed that the slide effect, visibility effect and the slide-visibility interaction all were significant at p = 0.0005, p = 0.0001, and p = 0.0099 respectively. The load effect did not show high significance (p = 0.113-1). Figure 43 shows the mean PDA as a function of the two levels of target visibility (nominal level). It can be noted from the figure that the PDA corresponding to Slide 4 is different from those of Slides 1, 2, and 3. The reason for the above effect will be obvious if one reviews the visibility data under variac settings 40 and 60 for all the slides. Within each variac setting let the slides be rank ordered in the increasing order of target visibility values in Table 25. One can observe that the above ordering of the slides is the same as the one produced by the PDA data within each variac setting. This result suggests that the PDA is directly affected by the target visibility value and not the nature of the images in the slides. Hence, the slide image may correspond to any highway target but the PDA produced in the laboratory during the testing of a subject will depend on the target visibility produced by that slide. 142
Table 25
Target Size and the Visibility Details for the Four Slides used in Stage IV Study
Details Slide 1 Slide 2 Slide 3 Slide 4
1. Vehicle Type White Dodge Dim Yellow VW Tan GM Van Half White VW
2. Vertical Visual Angle Subtended
at the eye - C> ' i—i Body 00 23° 26° 20° F ront 12° 14° 16° 12°
3. Luminance by Target Only (foot-Lamberts) V ariac 40 0.018 0.015 0.019 0.045 V ariac 60 0.200 0.100 0.120 0. 260
4. Luminance of the Scene (ft-L) 0.155 0.155 0.155 0.155
5. Target Visibility Variac 40 0.116 0.097 0.122 0.290 Variac 60 1.290 0.645 0.774 1.677
Note: The luminance of the target images was directly measured with the SPECTRA Photometer Model FC-200.
Image Luminance + Scene Luminance - Scene Luminance Target Visibility - Scene Luminance 143
1 0 0 - i Note: 1, 2, 3, and 4 refer to different slides. Values within parentheses refer to target visibility of a slide under given variac settings.
90 (1.290) (1.677) (0.774)
(0.645)
(0.290)
8 0 -
(0 . 122)
(0.116)
70- (0.097)
40 60 Variac Setting
Figure 43— Mean PDA as a Function of Variac Settings for Four Different Targets (Slides) Pooled over Two Loads. (Subject S03). The data on PDA was transformed into PDAT as discussed in the Stage III experiment and the residual plots were obtained for both PDA and PDAT. The ANOVA tables from the regression analysis and the residual plots are given in Table 41 and Figures 69 and 70 in Appendix A. From the ANOVA Table for the transformed variable PDAT it can be seen that the nature of results are the same as for PDA. Also residual plots do not show any special trend.
Experiments in Peripheral Detection
Experiment 1:
Purpose:
The objective of this experiment was to investigate the effect of three foveal task loads and four target visibility levels on the peripheral detection angle of two subjects. The selection of the levels of foveal task loads and the target visibility was to be based on the individual subject’s critical "on time” and threshold visibility data.
Design:
Foveal task load and target visibility were the two dependent variables in the experiment. Figure 44 shows the design for the experiment.
Session 1 Session 2 Session 3 F. Load 1 F. Load 2 F. Load 3
*1 *2 fcl *2 t l *2 *-> & 0) •H|M hfl 1 3 3 4 4 3 2 t-< a 2 4 12 2 14 ri •HCO H > 3 1 4 3 1 4 3 ______4 2 2 13 2 1 t and t represent the time period within each session
The numbers in each column refer to the order in which each of the four target visibility levels was presented to the subject. Number of subjects tested - 2. Number of replications - 2.
Figure 44.—Experimental Design for Visual Field Experiment 1. 145 The foveal task load levels for each subject were determined in the following manner: It was decided to make the critical "on time" as the second level of foveal task load. This load was considered to be stressing the subject just to the point where the subject can perform the foveal task without commit ting reasonably large number of errors (two or less per 20 trials has been considered to be the limit). The first level of foveal load was obtained by adding 1000 msecs. to the critical "on time". The third level of foveal load was obtained by decreasing the critical "on time" by 250 msecs. Thus the third level of foveal load was purposely designed to force the subject to cause errors in the foveal task. Hence, one of the secondary objectives of this experiment was to determine the effect foveal loads of the third level on the peripheral detection angle (PDA).
The levels of the target visibility were based on the subject's average threshold visibility. The following four levels were fixed for the target visibility:
Level 1: 0.5 threshold Level 2: 1.0 threshold Level 3: 2.0 threshold Level 4: 4.0 threshold
P rocedure:
The procedure for data collection was the same as described in the pilot experiments. Including the replications each subject was tested for six sessions. Each session consisted of eight observations after which the subject was given three minutes of rest.
R esults:
Peripheral Detection Angle:
The data on the PDA was analyzed using the ANOVA model with subject and replication as random factors and load visibility as fixed factors. The ANOVA Table is given in Table 42 in Appendix A. It can be seen from the ANOVA Table that visibility effect was significant at p = 0.0014. The average PDA at each foveal load has been plotted for the four visibility levels in Figure 45. It can be seen from the figure that the effect due to the subthreshold visibility can be clearly distinguished from the effects of the other visibility levels. A contrast was constructed to study the difference between subthresh old visibility versus the rest of the three visibility levels. The sum of squares due to visibility can be partitioned into sum of squares for the contrasts subthreshold and the remainder as follows: Figure 45. —Mean Pda at Four Levels of C ontrast as a Function of Three Levels Levels Three of Function a as ontrast C of Levels Four at Pda —Mean 45. Figure
Mean PDA from Straight Ahead (Degrees) 100 0 “ 90 “ 0 8 of Foveal Load and Probability of E rro r and Response Overlap in in Overlap Response and r rro E of Probability and Load Foveal of oel aka Fnto fFva odfrSbet SOI& S02 Subjects for Load Foveal of Function a as Task Foveal Combined. 3 2. (Threshold) .0 2 - C3 4- 40 (Threshold) 4.0 - C4 . (Threshold) 0.5 - isibility l V C Threshold - C2 oel ak Load Task Foveal P eripheral Task eripheral P Foveal Task Foveal MediumLow Probability of Response Response of Probability Overlap 'robability 'robability C4 C2 fEror rro E of l C High 0.2 146 ■8 ■s 147 Effect ■Sum of Squares DF MS'
Visibility 1219.44 3 406.48
Subthreshold vs the other three levels 1166.00 1 1166.00
Rem ainder 53.44 2 26.72
It can be seen that the contribution by the subthreshold visibility is much larger compared to the remainder. Hence, for subsequent experimentation it was decided to use only three levels of visibility. The higher level, i.e ., 4.0 (threshold) was dropped.
The variable PDAT was also analyzed in the same way as PDA and it can be seen from the ANOVA Table (Table 44) in Appendix A that the nature of the results are not changed.
The mean angular speed of the target was maintained at 3.05 degrees per second with a standard deviation of 0.53 degrees per second.
A regression analysis was carried for PDA and PDAT with speed as a covariable and the rest of the variables the same as in the ANOVA model. The regression coefficient for speed was not significant for PDA, but for PDAT it was significant at p ^0.1.
The residual plots from the regression analysis are shown in Figures 71 and 72 in Appendix A. No trend was observed in the residual plots suggesting that the re ression model was adequate.
Foveal Task Errors:
From the foveal task response two performance measures were computed:
Probability of Error = _____ Total Number of Errors _____ Number of Trials on the Foveal Task during one visual field observation
Probability of Response Overlap: This corresponds to the number of times response to a given stimulus was made in the cycle time of the next stimulus divided by the total number of trials. 148 These measures have been plotted in Figure 45 for the three levels of loads. It can be seen from the figure that the probability of error was negligible at low foveal load. At medium foveal load one would expect the probability of error to be less than or equal to 0.1 by definition of "critical on time". From the figure it can be seen that the probability of error was actually about 0.065. At extreme foveal loads the probability of error increased sharply to 0.245, but peripheral performance did not suffer as much as foveal task performance. Hence, a second important conclusion from this experiment was that at extreme foveal loads the performance decrement can be observed only in the foveal task.
The foveal task performance measures reported in Figure 45 corre spond to a situation in which the subject was performing both the foveal and peripheral tasks. The first two levels of foveal task were designed such that the subject can maintain a performance level with no more than two errors. It is possible for subjects to produce a reduced level of foveal performance in the presence of the peripheral task. Hence, foveal performance data without peripheral task were collected for subjects SOI and S02. This data is summa rized in Table 26. From the table the following can be observed:
1. For foveal loads with display "on time" equal to or greater i than the critical "on time" mean probability of committing errors within a sequence of trials is less than the performance level criteria (0.1).
, 2. For foveal load with display "on time" less than critical "on time: (extreme loads) the mean probability of error occurrence is greater than the level of performance criteria.
Hence, it can be concluded that within the critical foveal load the presence of peripheral task does not affect the subject's foveal task performance.
Experiment 2:
P urpose:
The purpose of this experiment was to determine the effect of three levels of target visibility and three levels of foveal task load on the PDA of three young and three aged subjects.
Design:
Three test sessions were planned for each subject. During each session tests were planned with one foveal load condition. Hence, a 3 x 3 Latin square design was followed as shown in Figure 46 with the three subjects and three 149
Table 26
Foveal Task Performance Data With and Without Peripheral Task for Subjects SOI and S02
Display Subject Probability of Errors Probability of Overlaps With Without With Without On Time msecs P-Task P-Task P-Task P-Task
1750 SOI .0126 .0081 .0000 .0000
750 .0429 .0175 .0122 .0000
500 .2179 .1871 .0200 .0420
2000 S02 .0069 .0000 .0000 .0000
1000 .0829 .0858 .0563 .0277
750 .2665 »1000 .0875 .1393 150
Sessions
1 2 3
1 L3 L2 LI L I C ritical on tim e + 1000 m secs Subjects 2 L2 LI L3 L2 Critical on time L3 Critical on time - 250 msecs 3 LI L3 L2
Each cell includes 9 observations that were made by changing three levels of target visibility.
Figure 46. —Latin Square Design for Testing Three Subjects During Three Sessions with Three Foveal Loads.
Periods Periods Periods
$ 1 0) ©► >> 2
S3• 1 -4 •HCO >
Assigned to LI Assigned to L2 Assigned to LI
Figure 47.—Latin Square Arrangement for Ordering the Observations within a Period Under Three Visibility Levels. 151 sessions as the rows and columns of the square and the foveal loads as the cells of the square. During any test session 9 observations had to be made by chang ing three levels of target visibility. These 9 observations were considered as being made in three successive time periods. During each period all the three visibility levels were to be used in the test conditions. To balance the order of presentation of the visibility levels across foveal loads and time periods three different Latin square designs were considered with the visibility levels as rows, time periods as columns and order within a period as cell numbers. These are shown in Figure 47. The three levels of foveal load are as follows:
Level 1: Critical "on time" + 1000 msecs. Level 2: Critical "on time" Level 3: Critical "on time" + 250 msecs.
The three levels of target visibility are as follows:
Level 1: 0.5 x Threshold Visibility Level 2: 1.0 x Threshold Visibility Level 3: 2.0 x Threshold Visibility
P rocedure:
The testing procedure was the same as in Experiment 1.
R esu lts:
Peripheral Detection Angle:
The results on peripheral detection angle (PDA) were analyzed in the following three steps:
Step 1. The data for each age group was analyzed separately using the ANOVA Model with the following main effects and interaction effects:
Subjects Foveal Load Target Visibility Session Time Period within a Session Interaction of Foveal Load and Target Visibility Residual
The ANOVA tables for young and aged group of subjects are given in Tables 46 and 47 in Appendix A. The analysis showed 152 significant subject, and contrast effects for both the groups. (Subject Effect - p = 0.0001 for both the groups, Contrast Effect - p = 0.0001 for both the groups) The load effect was significant at p = 0.09 for the young group and at p = 0.004 for the aged. The mean PDA for each age group by foveal loads are shown in Figures 48 and 49. From the figures it can be noted that both groups of subjects show a decline in the peripheral detection angle when the target visibility was reduced from twice the threshold level to half its value.
Step 2. In this step the data for both the groups were combined and analyzed. The ANOVA model consisted of the following main and interaction effects:
Age Subjects within Age Foveal Load Target Visibility Load x Target Visibility Age x Foveal Load Age x Target Visibility Age x Foveal Load x Target Visibility Test Session Time Period Residual
The ANOVA Table from Step 2 analysis is given in Table 48 in Appendix A. From the ANOVA Table it can be noted that Age-Foveal Load interaction was significant at p = 0.0006. Refering back to Figures 48 and 49, it can be seen that the aged subjects show a nonlinear relationship in their PDA under the three foveal loads.
Step 3. In this step all the main and interaction effects used in Step 2 were used in a regression analysis with target angular speed as a covariable. The ANOVA from the analysis is given in Table 49 in Appendix A. From the analysis the regression coefficient for speed was found to be significant at p = 0.0118. The result indicated a decreasing trend for the PDA with the increasing speed. The equation for the PDA in terms of an angular speed (0) in degrees per second can be stated as
PDA = 93.9 - 0.950 100 153 Peripheral Task
m © 0 u W>© P C3 T3 a? rC < C2 1•H d 90 u m B o S-t0 C l s P4 al0
80 C2 - Threshold Visibility Cl - 0.5 (Threshold) C3 - 2.0 (Threshold) 0.3 Subjects S07, SO 8, and SO 9 (Aged Subjects)
Foveal Task 0.2
'robability of Error 0.1 Probability
Probability of Response Overlap
Low Medium High Foveal Task Load Figure 48. —Mean PDA at Three Levels of Contrast as a Function of Three Levels of Foveal Load and Probability of Error and Response Overlap in the Foveal Task as a Function of Foveal Load. 100 154 Peripheral Thsk
C3
C2 T3
C l 90
Xfl
C2 - Threshold Visibility Cl - 0.5 (Threshold) C3 - 2.0 (Threshold)
80 - Subjects S04, S05, and S06 (Young Subjects) 0.3
Foveal Task 0.2 Probability Probability of Error 0.1
robability of Response Overlap
Low Medium High Foveal Task Load Figure 49. —Mean PDA at Three Levels of Contrast as a Function of Three Levels of Foveal Load and Probability of Error and Response Overlap in the Foveal Task as a Function of Foveal Load. 155 Figure 50 shows the plot of all the PDA data against corre sponding speed at which they were observed in the laboratory. The residual plot from the regression analysis is shown in Figure 73 in Appendix A. The residual plot does not show any specific trend indicating the goodness of fit of the regression model.
Foveal Task Errors
Figures 48 and 49 show the foveal task errors as a function of the three levels of foveal loads.
The probability of occurrence of errors increased approximately linearly with the foveal load for young subjects. At the medium foveal load the error was within the critical level of 0.1 as stated earlier under the criterion for critical load. The probability reached an average of . 15 under the high foveal load conditions.
The aged subjects showed an increase in the error probability beyond the critical level of 0.1 even in the medium foveal load conditions. This would suggest that the designed critical load is above the true critical load level for the aged subjects.
Visual Search in the Foveal Task
As reported in the literature search Bloomfield, Gould and Williams have investigated this topic in greater depth. Visual search time has been reported as one of the primary performance measures relating operator performance to task characteristics.
In this research the foveal task which the subjects performed involved visual search. The subjects searched the display to determine whether it contained "all 8's" or "one 6". This search task was performed under three different load conditions. A specific load condition was characterized by the display "on time". The loads were determined for individual subjects based on their "critical on time". Critical "on time"was defined earlier as the shortest display "on time" under which the subject could perform the search task with no more than two errors in twenty trials. This level of load is considered as critical load (or Medium load). The values of critical "on time" are given in Table 21. As described in earlier sections the low level of load for each subject was obtained by adding 1000 msecs. to the respective critical "on times". High level of load was obtained by subtracting 250 msecs. from critical "on time". The display "off time" was maintained as 250 msecs. Peripheral Detection Angle I CO.00000000 95.00000000 P0.00000G00 t f B .O O O O O O O O 7R eo.ocoooooo .00000000 iue5*Prpea DtcinAgev TagtAglrSed Eprmn 2) (Experiment Speed Angular arget T vs Angle Detection 50*—Peripheral Figure 1.00000000 » * . * T ' f I ■» ...B . 1.80000000 A ______-■ - A S A A aa AA BA A A A 6 AAA A A A A A B A B AA AA A A C A AA A A AB A A A A B
B A A A ______-- - - - A ______P E R I P H E R A LD E T E C T I O NE X P E R I M E N T2 - A BA B A A ' ______A A A A A
_____ A A ______...... A PLOT OF POA SPEEO VS A...._A B ______A 6 A A ... A . .6 agtAglr Speed Angular Target J» ______A A A AA 2.60000000 ___ A . A A. . . . . _ ..A . A...... __ . A . A....____ D A A AA A A AA A A A ...... A A AA . A A A A AA A A A _ A A ____ A _ . A A... A . . _____ 3.40000000 A ...... AC A ___ A t AA_ A A A B A ...... A A A 4.20000000 5.00000000 A A. 157 The data on total response times for correct responses was analyzed to obtain the following statistics:
1. Mean and standard deviation of response times for each subject for the three task loads
2. The above for only "all 8" condition of the display
3. The above for only "one 6" condition
It should be noted that the above response time data was collected when both the foveal search and peripheral detection task was performed by the subject. The data for the eight subjects who were tested in Experiments 1 and 2 are summarized in Table 27.
The data of the mean response time reported includes
1) an element of time related to search and
2) an element of time related to the response mechanism of the subject.
Hence, according to Bloomfield, mean search time under a given task load can be computed as shown below.
Mean search time Total response time Response time under given load conditions under given load under no search
The above methodology was followed to compute the mean search times for each subject for each task load. The data on the mean search times are also reported in Table 27.
From the tables it can be seen that the search times for "all 8" condi tions are generally greater than the search times for the "one 6" conditions. The reason for the above result is that during "all 8" condition subjects had to confirm each digit before making a response. But in the "one 6" condition the probability of detecting the presence of a 6 during the early part of the search e x ists.
Correlation Analysis:
This research investigated the effects of foveal loads and target visibility levels on the subjects' peripheral detection angle. It will be useful to determine the correlation between PDA and other variables such as subject Table 27
Mean and Standard Deviation of Response Times in a Foveal Task (in the presence of peripheral task) Both Types of Display
'COS SUPJECT-----RTL------'soRtr' RTM ~Sr>RTM‘ RTH SPRTT H”------sir STM 5"1h
1 SOI 850 234 t >14 15 5 __ _ 467 .....____ 153 544 ... . 308 ___ 161 2 S02 1037 282 846 257 791 24 8 601 410 355 3 SO 4 661 163 587 93 538 140 534 260 211 4 SO 5 1016 270 967 204 835 238 647 598 466 5 S06 1016 . . 264 . 1008. ?0 5 904 150. _____ 68 4_ ___ .676 __ __ 572_ 6 SO 7 946 280 771 225 826 27 2 566 391 446
7 SO 9 1499 429 1 1 2* ...?71 1141 _ ?J4 105 5 697 S S10 1079 241 924 201 69S 206 764 600 383
All-8’s Display
DBS'-----SUBJFCT “ RTL 8 “SDRTCB "RTFS------SDRTM8"...... RTH8 " SDRTH6 ‘ SI CP ‘ STH8 5TH8
1 SOI 1001 __ __ 1_6_5______650____ 157 461 _____ 1 58 _ 695 _ 344 155 2 S02 1136 241 888 271 BOR 243 700 452 372 3 S04 937 162 610 91 550 161 610 283 223 4 S05 1210 183 1062 167 . 894 235 841 693 525 5 S06 1156 214 1064 199 929 _ 136 824 732 597 6 S07 940 240 718 170 798 2 54 560 338 418 7 S09 1674 271 1229 247 1182 22C 1230 785 738 8 S10 1168 174 974 211 729 161 853 659 414
One-6 Display
"08 S SUBJFCT RTL'6" SDRTC6 RTFS SDRTM6 RTH6' 50RTH6 STC6 57TIS 3fTT?S~
1 SOI ______688______183 ...... 14 3 473 ____ 1 48______382 ___ 269___ 167 2 S02 920 284 800 235 772 2 52 484 364 336 3 S04 784 125 562 90 525 10 P 457 235 198 4 SOS 815 187 864 169 773 227 446 495 404 5 S06 868 229 939 193 875___ 160 536 607 543 6 . S07 952 341 842 268 860 290 572 462 480 7 S09 ______1330 ___ __ 487 1000 ______24 8 1 089 245 ___ 886__ 556 645 b S10 984 266 878 181 669 2 38 669 563 354 Note: RT, SDRT refer to response time and standard deviation of response times, L, M, H refer to Low, Medium, High foveal loads. 8 and 6 refer to display condition. 159 descriptions and baseline data. The basic assumption in the correlation analysis is that each observation is independent of others for a given variable. The following variables for the eight subjects who participated in Experiments 1 and 2 were used in determining the product-moment correlation coefficients:
1. Subject Descriptors A. AGE - in years B. HEIGHT - in inches
2. Baseline Data A. NVF1 - normal visual field (left temporal) for small size target B. NVF2 - normal visual field (left temporal) for larger size targ et C. SRT - mean simple reaction time in msecs. D. CRTON - critical "on time: in the foveal task E. THR90 - mean threshold visibility (contrast at 90 degree eccentricity F. THR60 - mean threshold visibility at 60 degree eccentricity
3. Peripheral Detection A. PDA - mean peripheral detection angle (overall) B. PDACL, PDACM, PDACH - mean PDA under three levels (low medium, high) of visibility C. PDALL, PDALM, PDALH - mean PDA under the three levels of fovealloads
4. Search Time A. STL, STM, STH - search time under low, medium, and high foveal loads
The correlation matrix obtained from the analysis is given in Table 28. The following observations can be made from the results on correlation between v ariab les:
1. Peripheral detection angle is negatively correlated with a) threshold of visibility at 60° (-0.838) b) critical "on time" for the foveal task (-0.70) c) age of subjects (-0.666)
2. Mean peripheral detection angle is positively correlated (0.717) with the normal visual field for small target determined in the Eye Clinic. Table 28
Matrix of Product Moment Correlation Coefficients between Subject Descriptors, Baseline Data, Foveal and Peripheral Task Performance Measures.
N * 8 CORRELATION COEFFICIENTS PROB > lft1 UNPER HO: RH0»0
CRTON PDALL ...... STL PDA ------THR60 ------PDALH —------STH ' " "NVFI “ ■ " RJVF2 " " ' ' STM...... ACE C.87252 6 -0.776003 0.720717 -0.666596 0.607007 -0.605958 0.595596 -0.583196 0.51559? 0.501772 0.CC52 0.073? 0.0A?7 0.0695 0.1088 0.1CS6 C.1763 0.1275 0.1905 0.2039 PDALM SPT HF1GHT TH»°0 -0.501595 0.3<>el01 0.351091 0.?C'9102 0.3251 0.3299 0.5662 0.6726
ppalm CRTON ACE SRT NVF2 PDA POALH THR90 PDALL STH HFIC-HT 0.671130 0.360518 0.351091 0.331973 0.319986 0.287779 0.187202 -0.179787 0.106316 -0.10553? C.0670 0.6171 0.58R2 0.5 753 0.5560 0.5050 0.6591 0.6715 0.7979 0.79 PC STK T HR 60 NVF1 STL -0.103798 0.075761 -0.052170 0.016615 0.8009 0.6 570 0.«178 0.°675
CRTON S1H PDALH PPALL TWRO0 PDA THB60 SRT STL ACE NVF1 * -0.FAEA27 -0.P3 1F *>7 0.798727 0.738596 -0.731290 0.717924" -0.705 063 -03670256 -0:6’509« '-0.562196’ ■" 0.00ft1 0.011:6 0.0173 0.0355 0.0383 0.0539 0.0596 0.0675 0.C8°0 0.1275 . ; STM PPALM NVF2 HEIGHT -0.522565 P.60?3°8 -0.175397 -0.057170 0.1»26 0.2760 0.676 ft 0.9178
THRfi 0 PDALL PDA ACE POALH CRTON SRT HEIGHT STM PDALM NVF2 0.702253 -0.60729Q -0.575538 0.51559? -0.5077 95 0.551855 0.538188 C.319986 —0.307P05 —C.?P9°05 O.OSC‘9 0.1066 C.1796 0.1505 ' 0.20 ?fi '■ — 0.2729“ ‘ 072775" 0.5560 '" C. 5371 " 0.5091 ' STL NVF1 STH THR90 0.1CCF7F -0.175397 “0.051553 0.073,8P C.669 7 C. 67 PR 0.6230 0.8557
... Nvpz . NVF1 THR 60 THR90 ------STH------STC* POALH PDALL ~AGE SRT -0.670356 0.61256ft 0.5F1752 0.57610 7 0.5769 51 0.566865 -0.518158 0.538188 -0.508175 0.398101 C .0675 0 . 10 39 6.1290 0.1315 0.1326 0.1511 0.1869 0.2775 0.3163 0.3299 SIM PDA HEIGHT PDALM 0.365135 -0.363665 0.3219?3 0.011778 0.622?0.6217 0.5753 0.9775' 160
Note: See text for the description of Variable Names, Table 28 (cont.) V 0 4 4 4 cm 30 X O a 1 • I { I 1 - * i Ui 0 »>• 0 •< X X _J X i ? ccir, i r A 03^4 r,o . c r.< . a A4 IA H A f-• r-«o o m C 4 0 r-*r* cm sf-rw o m tM 4 fyH 0 4 • • • • t • • 1 *- h ) N C V) W’.W £ VI fiC •OO' CL Ct < h* CL c X » o«n n « .o *o ! • o W o .m >- .4 i 1 • 4 r-oc 1 1* (AO o m ! • ' IAO •OH r- o C •“* C »-»cG. CO • •MH CM CM•AO- 1 O U T I O • «*> ■M o 0 -40 CM 0 a>0 I-• f-o 00 - f-o O N o* in cm 4 * o m tu«-i m-* 44 4 4 *0 \ u * • r-a> > a Q> *■■CM C O 3 HN < U m.t •< •< r»a r j t'l £ to r*; • 0 . O* /*.* 0 0 03 (MCI 0 4 0 «-*o r-r- f f JO o 4 • 4 st kT 0 "siO 1 1 • • o' cm c C * CC o ►- (A 4 - r O AO O (A VO m a or o« o Mcclu O . AO IA O U. lA CM a: ►- UJ - h X X a 0 X «v IT IT «v X • - r (ACL W'.OO m r « 0 00 4 CM •4 h (Si«0 r t 00 O • m - r • ^ O J 0 © m 0 0 U • (U 0 AO (A (7 OD (7 • • • • (O m O cm O 4 , 0 4 <4 H H < U . «Mm «Mm . U 0 • o 4 o ix o- *0 i * j • i i - o -o a *- 4 4 - I «-• tsi flC O X 1 © cm » , « X vu *MO r-m O x L 4.H . 4 LL. (M 4 ) / mm CM (A a 0 CMC o c O (A • CM o m o m o - m AHH (A 1 4 I** 4 4 •4 Xi*0 O• I » 1 : - • ia . H- . u CM - r W 4 (O m o O uli • ® a 40 O 2T: . > v> X < ©INI ©INI < |T,< 1 mN j if • f^o I ? : r*»H A-• m i/i H uc J K A* CM IMf*" •*0 jf* C N H M4 4 IM O C ( • o 0 CMO * o 10 0 m o a © 4 • 4 O-CM 0 4 • C A < 4 ' * 0 4 3 - 4 1 1 • a 161 Table 28 (cont.) PDA PDALL PDALH HEIGHT THR60 STL NVF1 AGE STH THR90 PDALM 0.B7245? 0.7c62 14 -'0.740535" 0.6 Tli5u —0.5375lv —0.4 041 —PTKOZSRIT -0.401594 ' —0.3E 78?6 —0.346812 0.0052 0.0204 0.0298 0.0670 0.1678 0.3217 0.2 240 0.3251 0.3440 0.5968 C? TOM NVF2 STM SRT -0.33B355 -0.289905 -0.217381 0.011228 0.5841. 0.5091 0.6088 0.9774 PDA PDALL THR6C NVF1 PDALH CRTON THR90 STL AGE STH PDALH 0.954404 0.91329! -0.906314 0.798727 0.7405 35 -0.719253 -0.697862 -0.662579 -0.605948 -0.5898 42 0.0005 0.0020 0.0025 0.0173 0.0348 0.0433 0.0530 0.0719 0.1096 0.1? 20 SRT NVF2 . STM HEIGHT —0.518158 ---0.5P2794 ‘-0.244518 ' “ 0.1E720Z 0.1869 C.2028 0.5639 0.6591 • . ------STH ------STI< AGE ------ppalh ------NVF1 " ' 'CRTON " ' 'THR90 PDA ------TRPSO SRT STL 0.796124 0.762643 0.720717 -0.662579 -0.635098 0.611000 0.598224 -0.597 392 0 . 57 8 2 3 4 0.576951 0.020*. 0.0272 0.0427 0.0719 0.0890 0.1058 0.1154 0.1161 0.1314 0.1326 PDALL PTALK NVF2 HEIGHT “0.576939 -0.404124 0.180628 0.016814 0. 1226 ------0.2217 — “ ~ P.6697 0.9675 STH STL NVF1 AGE CRTON THR90 SRT NVF2 PDALL PDA STM 0.862276 0.762643 -0.527464 0.5C177? 0.424316 0.366065 0.364135 -0.307805 -0.266721 -0.262340 0.0063 0.0272 0.1626 0.2C39 0.2951 0.62 52 0.6223 0.5371 C.527 8 0.5348 ~ "' PDALH ------PDALH------HEIGHT ” " THR60 -0.244518 -0.2173PI -0.103798 0.0857°fl 0.5639 O.feCPR 0.B0O9 0.8333 STH NVF1 STL CRTON PDALH AGE SRT PDA PDALL TWR90 —STH ------0.H62276 — -0.82 18 97 0.7661?4 0.70004 5 -0. 5898 4? 0. 5 84506 0.57PT97 — -0:55HR39'“ -0.557 504 015T9197 - 0.0063 0.0106 U.G2G4 0.0520 0.1220 0.1263 0.1315 0.1482 0.1494 C.1B5E PDALH T HR 60 HEIGHT MVF2 -0.387826 C.372975 -0.105432 -0.091453 0.3440 0.6352 0.7960 0.8230 Note: 1) Each row in the table represents the correlation coefficient and its significance probability between the ...... variable on theieft and other variables. ______2) The significance probability of a coefficient is the probability that is greater than or equal to the stated — value. 162 3) See text for the description of the variable names. 163 3. Search time in the foveal task under extreme loads was negatively correlated (-0.8318) with the normal visual field data (for small target) obtained from eye examination. 4. Search time under high foveal load is positively correlated (+0.70) with the subjects' critical "on time" for the foveal task. It should be noted that the search time under low and critical loads was not correlated with critical "on time". 5. Mean visibility thresholds at 60° and 90° were negatively correlated (-0.705 and -0.731) with the normal visual field data (NVF1). 6. It can be seen from the table that among the subject descriptor variables and baseline data normal visual field (NVF1), critical "on time" and visibility threshold at 60° were correlated with most of the performance measures. Correlation between PDA and Foveal Task Errors For each subject in Experiments 1 and 2 the data on peripheral detection angle, probability of occurrence of errors and probability of response overlap in foveal task were considered for determining product moment correlation coefficients. The correlation matrix by subject is given in Table 29. It can be seen from the results that the peripheral detection angle and the probability of occurrence of errors in the foveal task were not correlated each time for each subject. PDA was also not correlated with the overlaps in foveal task response for all the subjects. Also for most of the subjects response probability of occurrence over laps and errors were correlated (correlation coefficient approximately equal to 0.40). reainMarxfr D, re pe n oelTakErors rro E ask T Foveal and Speed arget T PDA, for atrix M orrelation C - - n e e n - • M © • I = 81 = N ' i rotv «ut oe 1 PO_ n PO_ poaiiyo curne of occurrence of probability PROB_E and PROB_0 1) Note: _SP_Et£i_ ------n PR0B_0 PR0B_E PDA -PR0B_E- — — PROR^F PR0B_0 PDA PE -0.352457 SPEED RB0~ -0.092796 ~ PR0B_0 PDA SPEED = ...... 23 2) Significance probability is given below correlation correlation below given is probability Significance 2) 153 — ■ — ------al ubjct CORRELATION COEFFICIENTS ts jec b su all a coefficient. response overlaps and e rro rs in foveal task . task foveal in rs rro e and overlaps response on s ect correl i ci s t n ie ic f f e o c n tio la e r r o c ts c je b su young __ ...... - .3C4 - 0.131C54 - - ...... -0.159704. ge ubjcts jec b su ed ag -0.266127 ..... 0.068652 000 -0.352457 1 .000000 1.000000 0. . 0 0 0 .0 -0 0.0000 - 0.0000 0.0001 071253 0.2334 000000 077863 0.2772 321833- 0.0000 0.0458 0.0001 .920.8136 0.5962 0.5850 0.2389 0.0155 PDA PDA PDA ______al 2 164 29 Table • —0.1724 •—0.1724 87— "-0.018679" -0.021704 0262 -0.092796 -0.266127 0181 0.276883 -0.128611 0.0155 - 0.148047 0.077863 1.000000 1.00 00 00 00 1.00 0.010259 1.000000 0001- 1 0 0 .0 0 0.2334- SPEED" 0.7459 0.0221 0.0000 RB0PROB_E PR0B_0 0.0310 0000 0 0 .0 0 0.9248 0.2512 0.0000 SPEED CORRELATION COEFFICIENTS ------0.665 0 0.665 ------. ------...... ------0.071253 -0.018679 0.148047 0.06B303 0.137019 1.000000 0.010259 -0.159704 0.068652 KOOOOOO l.oonoon - 2 7 7 2 . 0 .oooo 0 0.2984 0.0221 .830.0000 0.0873 5962 6 9 .5 0 0000 0 0 .0 0 .160.0310 0.8136 0.0000 0.0119 0.2512 0.9248 PR0B_0 PR0B_0 ---- ___ ---- ~ ~ __ ------1000 .... 1.000000 - 1000000- 0 0 0 0 0 —1.0 ------0.321833 -0.021204 ------0.128611 1.000000 0.068303 6^ 13701 9~’ 0.172487 .783 ~ 0.276883 0.131854 0. 1 0 0 .0 0 0.045B-- .39 — 0.2389 0.7459 0.0000 0.2984 O.OB73 0.0119 PR0B_E 0.0000 RB E ’ PROB_E SPEED ------Table 29 (cont.) 165 36 Subject 1 CORRELATION COEFFICIENTS J PDA SPEED PROB_E PR0B_0 o PDA 1.006000 -0.179681 -0.155711 0.007967 ------0.0000.... - - 0.2946- -...... -0.6326 ------0.962?- — SPEED -0.179681 1.000000 0.439965 0.264145 0.2946 0.0000 0.0072 0.1 159 prob T e" -0.155711 0.439965 1.000000 ~ 0.49230 6 0.6326. 0.0072 0.0000 0.0026 -PROBED------—0.007967 - — 0.266145 ------0.492386 ------1.000000— 0.9622 0 .U 5 9 0.0026 0.0 0 0 0 * N s 36 Subject 2 CORRELATION COEFFICIENTS *s PDA SPEED P R o O PRDB 0 I PDA 1.000000 -0.305087 -0.030923 0.390997 *------0.0000 - 0.0671 ------0.0521------0 .0 1 7 4 -- SPEED -0.305067 1.000000 0.049607 -0.292520 0.0671 0.0000 0.7708 0.0799 PROS'"! -07030923 07049607 l7oOOOOb ~0.246745 " 0.6521 0.7708 0.0000 0.1433 —RRjQB^O______0.390997____ -.0.2.92520___ —0.246745 ______1.000000_ 0.0174 0.0799 0.1433 0.0000 * N = 27 Subject 4 CORRELATION COEFFICIENTS : PDA SPEED P M i O PRCB_0 — PDA 1.000000 -0.038287 -0.270577 0.230221 ______0 .0 0 0 0 ------0.8437 - -0.1692 ------0 .2 4 6 7 - SPEED -0.038297 1.000000 0.203981 -0.076560 0.8437 0.0000 0.3062 0.7056 PROlTE -0.270577 0.203981 1.000000 0.431721 0.1692 0.3082 0.0000 0.0232 —P-ROB-O------0.230221 ---0.076560 ------0.431721 1.000000 — 0.2467 0.7056 0.0232 0.0000 27 Subject 5 CORRELATION COEFFICIENTS PDA SPEED PROB_E PR0B_0 i PDA 1 .000000 -0.410499 -0.285266 -0.268941 - - 0.0000 0.0316 - 0.1459 0.1719 SPEED -0.410 *9 9 1 .000000 -0.065550 0.2P8710 0.0316 0.0000 0.7299 0.1408 " pro V j : -0.285266 -0.069550 1.000000 0.411431 0.1459 0.7299 0.0000 0.0312 _ PR0B_O -0.268941 0.288710 0.411431 1 .0 0 0 0 0 0 0.1719 0.1408 0.0312 0 . 0 0 0 0 PDA ; PDA i Fbivtiii I rom iiii PDA | • —PROB^O- _PK0BJ3._ PROB ~ _E~ ------__£ROB O e N PE -0.117513 SPEED PR0B_E PROR 0 N PROB_E SPEED N = 27 27 = N SPEED PDA PROB_E SPEED a a » » "" 27 27 27 27 ...... ------...... Subject Subject Subject 7 Subject ujc 10 Subject 014 27 02 41 -0.1 -0.1'a?397 016?5 - 5 ^0*1269? Subject 9 Subject -0.355352 -0.174842 044 05 60 47 -0.4 0.154199 1 .000000 0.110251 0.161596 1.000000 1.000000 n.jo^M7 0.091076 PDA .000000 I - 0.5170 0.5659 0.5516 0, 0, 0.5900 0.0658 ft.riftfifi 0.5348 0.5741 0.6132 0000 0 0 .0 0 0.5105 0.6556 0.6135 0.01 62 0.01 n.nnon tfnari PDA D SPEED PDA correl i ci s t n ie ic f f e o c n tio la e r r o c 6 PDA . 0.5741 ... 0. - 3 5 8 4 1 .1 0 _ ------0. 257397 --0. 154059— 9 5 0 4 5 .1 0 - -0.117513 -0.111124 -0.080646 -0.355352 -0.447605 1.000000 1.000000 0.166213 0.032648 0.161596 1.000000 0.021749 1.000000 0162— 2 6 1 .0 0 -0.C658- 0.5874 0.5659— 0.1923 0.0000 0.0000 0.5512 0.6915 0.8669 0.0000 0.58 80 0.58 0.5748 0.9106 0.0000 SPEED SPEED SPEED al 9 (cont.) 29 Table CORRELATION COEFFICIENTS CORRELATION COEFFICIENTS CORRELATION COEFFICIENTS ------ ------____ ----- ... — 0.171 OP 1 -0.111124 -0.080646 -0.17484? __ 0.154199 1 .000000 0.110351 0.44075? —~ 0.44075? 0.399857 0.032648 1.000000 1.000000 0.391519 — 0.391519 0.6132 0.091076 060 - -0.6900 0.021749 1.000000 .6 . . 6 . . 0.561 0.5511 0.5874 0.0000 KB EPROB_Q PKOB_E R B EPR0B_0 PROB_E 0.0203 0.0000 0.6915 0.6659 prob 0000 0 0 .0 0 0.0367 0.655 6 — 0.655 0.0000 0.0411 0.9106 PROB_E L prob e ------___ —1.000000 —1.000000 ------1.000000 —1.000000 -0.121981 -0.257397 -0.132397 -0.12692 5 -0.12692 -0.141027 1.000000 __ 0.440752 0.154059 0.166213 0.399857 5105.. 5 0 1 .5 0 0.391519 0.103417 0.114853 0.0000 0.5511 0.1923 t Kft 1 7ft 0.0000 0.5512 0.5348 6135- 5 3 1 .6 0 0.0203 0.5880 0.0367 0.0000 0.0000 0.0411 0.5740 PR0B_O J o _ — ---- — STATISTICAL ANALYSIS SYSTEM PLOT OF PDA VS PROB_E Note: A, B, C, ..., H refer to -100.00000000 0 D __ ------0 D F A D B A ° 0 . . . . H- A—...... 0 ------0 B A P H H DMA H A E C HA HB 0 H F F ------c B A CC A 9 5.00000000 H © 8 ------c ------A-C -----E------H-C------A------G E F SP 0 O E c BA < t G B £ H A - - G - F B 6 CN ------C S : f E B 0 O G K G 0 E H B *p4 H A G E C H A o -90JJ0CC0CC0 ..AG G F E G C G F G F F F C H B © H F F 6 G g £ HB Q G A G FC H H H B H A e rt G 6 Li f~t ------A -G-- 0) B E H •4 £5.00000000 B C E E B Li E - - F 0) B C A- F A F B F AA 80.0 0000000 G fr G G 75.00000000 •♦r-WYAS ' 0 .0 1 0 0 0 0 0 0 0.11000000 0 .2 3 0 0 0 0 0 0 0 .3 5 0 0 0 0 0 0 0 .4 7 0 0 0 0 0 0 O.59000000 Probability of Occurrence of Errors in Foveal Task 167 Figure 51. —Plot of PDA vs Probability of Occurrence of Errors in Foyeal Task. Blank Page CHAPTER 6 SUMMARY OF RESULTS AND CONCLUSIONS This research primarily investigated the effects of foveal task demands and target visibility upon detection of highway targets in the visual periphery. The research was conducted in the laboratory where a moving highway visual scene was simulated to cover the subject's visual field. The entire research efforts can be divided into the following four phases: 1. The development of the laboratory set up 2. Baseline data on test subjects 3. Pilot experim ents on PDA 4. Major experiments on PDA Highlights from all the phases are briefly discussed in the following sections. Phase 1: Development of the Laboratory-Summary One of the major contributions of this research was the development of the instrumentation and the laboratory set-up to conduct research in the area of peripheral vision of drivers. The set-up simulated the visual scene of driving through films. It provided a minimum of 102° field of view on one side. Periph eral targets as passing vehicles can .be superposed on the moving scene. . The . luminances of the scene as well as the target can be controlled to produce dif ferent levels of contrasts or visibilities for the target. Target motion can be controlled to represent different relative velocities of passing vehicles on the highway. Different lane positions for the target can be developed. Different types of computer controlled foveal tasks can be generated to intensify the test subject's foveal search. Based on the information about the facilities (O'Hanlon, 1975) available for driver research, it can be stated that the set-up is unique and the first of its kind available for peripheral vision research in driving. Phase 2: Baseline Data on Test Subjects-Summary The following baseline data were collected on ten test subjects. 1. Normal visual field from eye examination 169 2. Individual subject's threshold of target visibility (for a given stationary highway target) at back ground luminance of 0.1529 ft. L. and target ec centricity 60° and 90°. 3. Simple reaction time 4. Critical on-time for responding to the 6 digit dis play with probability of error not exceeding 0.1. The above data were considered as baseline data for the following rea sons: 1. They are simple measures depicting an individual's foveal and peripheral capabilities. 2. The data gathering procedure can be easily repeated without elaborate special measuring systems. Normal Visual Field The normal visual field data are frequently reported in ophthalmic liter ature and hence no general interpretations will be made on the data obtained in this research. They were obtained mainly for purposes of correlation analysis with the performance measures used in this research. It was found from the correlation analysis that 1. The normal visual field data corresponding to the left temporal side for the small and large target were not correlated. 2. Only the visual field corresponding to the small target showed high correlations with critical on- time, search time, PDA, thresholds for target visibility, and simple reaction time (see Table 28). Hence, during the rest of the report normal visual field will be referred to the data corresponding to the left temporal side for small target, in the horizontal meridian. For the eight subjects tested in experiments 1 and 2, the normal visual field (NVF1) on the left temporal side in the horizontal meridian provided the following statistics: 171 Mean 51.75 degrees S.D. 13.14 degrees Min. 30.00 Max, 70.00 Target Size - Disk diameter 0.56 mm. Target Luminance - 100 abs. Background Luminance - 31.5 abs. Threshold for Target Visibility The threshold for target visibility or contrast was determined at 60° and 90° eccentricities with stationary targets. It should be pointed out that the size of the vehicle image was 12° horizontal and 16° vertical. The background luminance was 0.1529 ft. L. Summary of the threshold data can be given as follows: Threshold Visibility Threshold Visibility at 60° (THR60) at 90° (THR90) Mean 0.3825 Mean 1,3732 S.D. 0.1000 S.D. 0.3659 Min. 0.2330 Min. 0.8710 Max. 0.5600 Max. 1.9870 From the summary it can be seen that the mean threshold for visibility increases from 60° to 90° eccentricity. This type result is well established in opthalmic literature (NRC Committee on Vision, 1976, See Figure 12, Chapter 1). The target size for the data in Figure 12 is 10 minutes of arc. Rogers (1972) determined the threshold for a 5° stationary stimulus with 2.88 ft. L. background luminance to be 0.043 at 0°, 0.088 at 20°, 0. 092 at 40° and 0.142 at 55°. The importance of the threshold data in this research lies in that the targets used were of a much larger size which is almost equal to the size of real highway targets. The important use of the threshold data in the research was to deter mine the levels of target visibility for each subject to be used during testing in the peripheral task. A second important result from threshold for visibility experiments is the one related to the effect of partial and full exposure of vehicle image at 60° eccentricity. Under partial target exposure the target size was 12° x 16°. Under full target exposure, the size of the body of the vehicle, 30° x 24°, is 172 added to the partial target (front hood of the car)„ The results showed that the mean threshold decreased from 0.404 to 0. 374 when the target exposure area was increased from 192 sq„ degrees to 912 sq. degrees. The above result im plies that under background of luminance 0. 1529 ft. L., target luminance re quired for detection under partial exposure is 0.2146 ft. L. and the luminance required under full exposure is 0.2100 ft. L. It can be seen that the two average luminances stated above are not practically different from each other. Hence, in this research the difference between partial and full target exposure was con sidered negligible. Literature about the effect of increased target area has been discussed under spatial summation in Chapter 1. The literature states that beyond Rieco's area for summation, the intensity of the target alone is re sponsible for detection. This suggests that in highway driving, detection of ve hicles passing in the adjacent lane may not be affected by size variations in the vehicles since the vehicle sizes in the adjacent lane can range from 15 to more than 35 vertical visual degrees. The mean threshold data THR90 and THR60 were correlated with each other (see Table 28). Both measures showed high correlations with PDA, and normal visual field data. Simple Reaction Time The simple reaction time and thumb travel time were recorded for each subject. The data is summarized below by age: Total Response Time Reaction Time including (msec s.) Reaction Time (m se c s.) Mean S.D. Mean S.D. Young 235 54 349 54 Aged 275 55 393 72 The aged tend to have an increased reaction and response times of about 50 m secs. The data on response time for each subject was used in computing search times in the foveal task. Reaction time was negatively correlated only with the normal visual field (NVFl). Critical On-Time (CRTON) By definition this refers to the on-time of the display with 250 msecs. off-time at which a subject could perform with no more than 2 errors out of 20 173 trials without the presence of the peripheral task. The determination of CRTON may be considered as finding some form of limit to a subject's capacity in per forming the display task. According to the loading task paradigm suggested by Knowles (1963), Rolf (1972), and Senders (1970), the foveal display task maybe considered as a loading task designed to determine the effect of some indepen dent variables on PDA. The methodology followed to obtain critical foveal load served the pur pose very well. The subjects under this load were able to perform the foveal task and keep the errors within 10 percent. Also, it was pointed out during the discussion of results of experiment 1, that the presence of the peripheral task did not alter the performance in the foveal task. Under these conditions the foveal task became a loading task. Results from Pilot Experiments-Summaiy The following were the objectives of the pilot experiments: 1. To find the effect of repeated observations on the peripheral detection angle (PDA) 2. To find the effect of testing on two different days on PDA 3. To determine the effect of different foveal loads and target visibility levels on PDA 4. To determine the effect of different target images on PDA The overall objective of the pilot experimentation was to guide the selection of target visibility levels for the major experiments in PDA. A summary of the major results from the pilot experiments are given in Table 30. Results from Major Experiments-Summary The objective of the major experiments was to determine the effect of three levels of visibility and three levels of foveal loads on PDA. In experiment 1, one additional level of target visibility was used. The target visibility levels used in the experiments were based on each subject's threshold for visibility at 90° eccentricity. They were as follows: Low visibility - 0.5 (Threshold) Medium visibility - Threshold High visibility - 2.0 (Threshold) 174 Table 30 Summary of Results from Pilot Experiments Test Conditions Results Support 1. Target Visibility = 7.4 Repeated observations with See ANOVA Scene Luminance = _ in a given day did not affect tables 36 & 37 Target Speed: 0.1529 ft. L. PDA. (Appendix A) Mean = 3.16 ° /se c . for "TRIAL11 S.D. = 0.80 ° / sec. effect. Stage I and H. Same as 1. Mean PDA was not affected See Table 36 due to testing on two differ and Figure 36 ent days. (950 93°, 96.96°) Same as 1. Variability in PDA data was See Figure 36 found to decrease slightly Stage I. on the second day. Same as 1. plus Mean PDA \kras not affected See Tables 24 foveal load 0 & 2 bits by foveal load, but the cu and 39 . per second. mulative distribution of See Figure 37 PDA show an increased per cent of trials with detection below 97° eccentricity. Stage n. Target Visibility PDA decreased with de See Table 38 Low = 0.2943 crease in the visibility in Appendix A Medium = 0.7848 (from high to low 6° de and High = 1.5042 crease). Scene Luminance = PDA decreased only at 1.5 Figure 39 0.1529 ft. L. bits/sec. load. Foveal Load = (0, 1.2 and PDA decreased by about 9° Stage HI 1.5 bits/sec.) from low load-high visibil Target Speed: ity to high load-low visibil Mean = 3.12 °/sec. ity conditions. S.D. =0.63 ° /se c . Four different targets PDA was not affected by See Table 40 (see Table ). fovealload. in Appendix A Visibility ranged from PDA was directly related and Figure 43 0.097 to 1.677 to visibility levels of tar Stage IV. Scene Luminance, = n c. T gets and not to the nature 0.1529 ft. L. Foveal Loads-0.857 and of targets. 1.5 bits/sec. 175 Similarly, the foveal loads were based on each subject's critical foveal load. Low load - Critical on-time + 1000 msecs. Medium load - Critical on-time High load - Critical on-time - 250 msecs„ At the high load, errors in the foveal task were inevitable. Six subjects (5 young and 3 aged) were tested. The summary of results are provided in Table 31. Table 31 Summary of Results from Major Experiments Result Support 1. Mean PDA was significantly low er under sub See Figures 45, threshold visibility levels when compared with 48, and 49 and threshold and superthreshold levels. T ables42 and 48 in Appendix A 2. Mean PDA was not affected by foveal load. Experiments 1 and 2 3. Mean PDA for the aged was 89° as compared to 92° for the young. 4. Minimum value of PDA observed for the aged was 74° and for the young 80.5°. 5. The overall standard deviation of PDA was about 5°. 6. In the foveal task performance, the mean probabil Figures 48 ity of occurrence of errors were maintained within and 49 the critical level of 0.1 for the young subjects at Experiments low and medium loads. For the aged subjects, the 1 and 2 critical error rate was exceeded at medium load level. 7. At high loads, the foveal task errors went up as Figures 48 expected, but did not significantly affect the PDA, and 49 and Table 48 Experiments 1 and 2 Discussions 176 Effect of Foveal Task Loads One of the major findings of this research is that foveal task loads sim ulating active search within a foveal region of 6° x 3° does not seem to have an appreciable effect upon the detection of vehicles passing from behind (see Figure 52). Previous research in the laboratory by Gasson and Peters (1965), Huntley (1973) Kephart and Chandler (1956) and Moskowitz et. al. (1972, 1974) concluded a possible reduction in the visual field with increases in foveal or central task loads. This conclusion is not supported by this research. Two important differences can be cited between this research and the previously re ported literature. 1. In this research, the foveal task loads were related to the limit of the individual subject's capacity. Where as in the previous studies, such a relation was not es tablished, and hence, those results are confounded with the subject's spare capacity. 2. In this research, the foveal loads provided loadings to the subject's visual search activity only. Efforts were made to eliminate motor control processes from the visual information acquisition process. But in the previous research, foveal tasks involved tracking, mental manipulation, etc. in addition to visual perfor mance. It should be pointed out here that Kochhar (1974) found that increased driving speeds in a simulator did not affect the response times to peripheral signals. In Bhise's (1971) work, it was reported that at higher eccentricity angles the performance in the peripheral task, either signal detection or car following tasks, showed degradation more quickly than the foveal task. This result does not agree with the results of this research because Bhise's subjects had to per form one of the following in their visual periphery: 1. Discriminate the position of the Landolt ring or 2, Judge the changes in the motion of the lead car. The above tasks either involve resolution threshold or threshold for judging the changes in the motion in the periphery. But in this research, de tection threshold is involved. Cumulative Percent of Trials in which H Q ■*-> ct 0 tf e> to 0) Figure 52. —Cumulative Distribution of PDA for Three Three for PDA of Distribution —Cumulative 52. Figure too 30 10 40 50 70 SO 90 70 Target angular position in degrees from straight ahead straight from degrees in position angular Target ifrn Fva Las Eprmns &2) 1 (Experiments Loads Foveal Different 58 9580 85 75 Load Low Load High Medium Load 90 for each curve each for 177 178 The lack of dependence between foveal and peripheral performance is also supported by the lack of correlation between PDA and foveal task errors (see Table 29 )„ Effect of Target Visibility Target visibility was found to affect PDA more than any other variable. Target visibility in this research is equivalent to target contrast reported in the literature. From Figure 12 it can be seen that contrast thresholds depend on background luminance. In this research, the background luminance was kept constant at 0.1529 ft. L. throughout the investigations. The results of this research with respect to target visibility are unique and not found in the literature because of the size of the target involved. It should be emphasized that the target size used in this research is approximately equal to the sizes of highway targets in the driving environment. From threshold visibility studies, evidence exists to suggest that when dealing with large size targets in the visual periphery, differences in size may not affect PDA, but target luminance or contrast will be the factor affecting pe ripheral detection, Ricco's law on spatial summation states this information, but literature does not provide any research evidence to this effect. At this point it may be appropriate to speculate that large target sizes may wash out the effects of foveal task demands. It is also observed that peripheral detection suffers most when the visi bility levels are below threshold level (see Figure 53). This implies that in driving conditions such as dawn and dusk with dark color cars, may lead to late detection. The background luminance (0.1529 ft. L .) used in this research falls within the mesopic range (log mL. 1 to -3). From Figure 53, it can be seen that it is reasonable to assume that pass ing vehicles with a target contrast equal to 50% of subject's threshold visibility at 90°, and background luminance equal to 0.1529 ft„ L ., will be detected by the subjects before the vehicle reaches 74° eccentricity. Effect of Age It is found from this research that the three, aged subjects showed higher target visibility thresholds, higher critical dn-time for the display task, in creased reaction time, and reduced PDA (see Figure 54) during most of the testing conditions. Burg (1968), and Wolf (1962, 1967) report the shrinkage of lateral visual field due to age. 179 90 for each curve o SO Cl Hr o C2 o c® o w GO a> rt a * 40 Visibility Levels C 3 P< to j» £ 3 0 L Cl = 0 . 5-Threshold 20 C2 Threshold C3 = 2*Thresholds 10 70 75 80 85 90 95 105 Target angular position in degrees from straight ahead position Figure 53. —Cumulative Distribution of PDA by Target Visibility Levels (Experiments 1 and 2) too Young 90 Young 5 Subjects N= 51 70 Aged 3 Subjects N= 27 40 £ H Aged -Young 10 70 7580 85 90 95 Target angular location in degrees from straight ahead Figure 54. —Cumulative Distribution of PDA by Age Groups for the Low Visibility Conditions (Data from Experiments 1 and 2) 180 Effect of Speed Target speed was not specifically investigated in this research. Litera ture on the effect of target motion on peripheral performance is limited (Low, 1947; Rogers, 1972; Leibowitz and Johnson, 1972; McColgin, 1960). As r e ported in Chapter 2, Graham (1967, 1971) concludes that evidence regarding the effects of large size moving stimuli on subjects’ peripheral performance is not available from the current literature. In this respect, this research sug gests some surprising results. The target angular speed during experiments 1 and 2 were maintained at an average speed of 2.85 degrees/sec. with 0. 75 degrees/sec. as standard de viation. Using speed as a covariable, it was found from regression analysis that higher target speeds lead to slightly reduced PDA. This is also supported by the correlation analysis. Figure 55 shows this trend about speed. Correlation of Performance Measures with Baseline Data From the Table of Correlation Matrix in Chapter 5, it can be seen that peripheral detection angle correlates well with the following three baseline data: 1. Threshold visibility at 60° 2. Normal visual field 3. Critical on-time for the foveal task It should be noted that this correlation supports the basic notion in the research methodology of this research i. e. levels of foveal task loads and target visibil ity should be based on individual subject’s capacity limitations. Search times in the foveal task are correlated negatively with the normal visual field data. This result is supported by Johnston (1965) and Erickson (1964). The increase in search time with the increase in total uncertainty ordis- play size shown in Figure 34 has been well established in the visual search lit erature (see Bloomfield, 1973). In conclusion, this research has provided interesting insights into the mechanism of peripheral detection of passing highway vehicles. Target contrast, motion, and size play an important role in determining a subject’s peripheral detection angle. Only a few aspects of these variables were explored in this research. It has opened new doors for research in this area. Some of the sug gestions for future research are described in the next section. Peripheral Detection Angle lO C .O O C C C O O O S 0.0 0000000 95.00000000 75.000C30C3 $5.00000000 e 0 .oooooooo Figure 55.—Plot of PDA vs T arget Speed for the 8 Subjects in Experiments 1 and 2. and 1 Experiments in Subjects 8 the for Speed arget T vs PDA of 55.—Plot Figure I.COCO0000 S T A T I S T I C A LAN A L Y S ISY S S T E M 0 0 1.80000000 t>D 0 ISC 0 CE 8 D H 0 H H------0 C B OH 0 - F_ H G- H E -.G ' P L O TO PD F AVSP S E E D b B HO -G 0 C G G A T arget Angular Speed in degrees/sec in Speed Angular arget T — H—8-0—H B ------EH 2.60000000 c A _c— HE H F H 0 G F EG C A H C _ C EC-a_C FH A D F ...... C E C B 0 E F ~F a a ------ A_C A AE A C G G G E B C E A G BA DE ------BA B H h ------AA AC C B Bft B a . O O O O4.20000000 3.AOOOOOOO C GGH C B F H - 0 e B—AH—. F HF F — F H FFB B F 0 B 88 F B ------AA- A 8 BE -AFt G- F G 8 ------B ------______Note: A, B, C, C, B, A, Note: Subjects l r 2f 4, 4, 2f r l Subjects ..., 5.00000000 H refer to to refer H ., 30. ..., ■ 181 Suggestions for Future Research 182 The following are some of the primary suggestions for future research, 1. Since there is some evidence from this research that higher target speeds lead to reduced peripheral detection angle, target speeds higher than 3° may be selected and the effects of these speeds on PDA can be determined. 2. The relationship between contrast thresholds and target ec centricity at different background luminances are available in the literature for small targets. For large size targets, experimental evidence regarding such relationships are not available. Data on contrast or visibility thresholds as a function of target eccentricity for targets of sizes com parable to highway targets can be directly used by designers of automobiles and highways in their determination of visual field requirements for the driver. 3. The laboratory set up can be easily adopted to determine the effect of eye-head geometry on peripheral detection angle for targets presented in the far periphery. 4. The foveal task can be made more realistic, but restricted to visual processes only by simulating a sign reading task. 5. The present slide projection set up can be modified by using a mechanically operated zoom lens to provide a simulated lead car and its motion. With this arrangement, thresholds for detection of lead car motion under different background luminances can be investigated. 6. The most interesting research that can be done is the investi gation of the leading edge concept in motion detection. Rec tangular targets of different width and height comparable to highway vehicle sizes can be introduced inward from the far periphery. The detection thresholds, for target contrast can be investigated under different background luminances for different levels of target motion. Finally, it is to be concluded that research in the area of human periph eral vision can provide useful information about human perceptual capa bilities. Most literature in this area have used small size targets and found significance with foveal task loading. But this research has provided initial evidence that large target sizes, such as those encountered in driving, wash out the effect of foveal attentional tasks. Hence, further detailed investigations are necessary to provide improved knowledge in this area. APPENDIX A ANOVA TABLES AND RESIDUAL PLOTS 183 Table 32 ______REACTION J.IHE. FOR_T6N .SUBJECTS______ANALYSIS OF VARIANCE TABLE • REGRESSION COEFFICIENTS . AND STATISTICS OF FIT FOR OEPENOENT VARIABLE REACTION SOURCE DF SUM OF SQUARES MEAN SQUARE F VALUE PRCB > F R-SQUARE C.V. P.E G R f S SI ON______307040.9233 0640 ...... 56337.88036738 ______33.06847_ .0.0001 ______0.51258684 16.41409 • FPP.OR 283 482X29.60911681 1703.67353045 REACTION KEAN CORRECTED TOTAL 292 989180.53242321 __ ST0 DEV ...... "...... ‘41.27558032 251.45051 SOURCE OF SEQUENTIAL SS F VALUE PROB > F PARTIAL SS F VALUE *»POB > F AGE 1 112695.56296905 66.14857 0.0001 113602.88698991 66.68114 0.0001 — SUBJECT!AGE) ------T8 394345.36033735 ---- 28.93346 0.0001 394345.36033735 28.93346 0 .0 0 0 1 184 Table 33 reaction tine .for . ten .subjects ANALYSIS OF VARIANCE TABLE t REGRESSION COEFFICIENTS * ANO STATISTICS OF FIT FOR DEPENDENT VARIABLE TRAVEL SOURCE OF SUM CF SQUARES MEAN SQUARE F VALUE PROB > F R-SQUARE C.V. REGRESSION ______9 105392.38621991 11710.26535777 22.35980 0.0001 . 0.41557684 .. 19.81818 t. ERROR 263 148212.66980057 523.7196B127 STD DEV TRAVEL MEAN CORRECTED TOTAL 292 253605.05802048 '22.88492258 115.47440 SOURCE OF SEQUENTIAL SS F VALUE PROB > F PARTIAL SS F VALUE PRCB > F AGE 1 1354.14451038 _ 2.58563 0.1090 1370.94410335 2.61771 . . 0.1068 _ SU5JFCT(AGE) 8 104038.2*370953 24.83157 0 .0 0 0 1 104038.24370953 24.83157 0.0001 Table 34 ______REACT ION.TIME ..FOR_.TEN.SUBJtCTS ______... ____ ANALYSIS OF VARIANCE TABLE « REGRESSION COEFFICIENTS * AND STATISTICS OF FIT FOR DEPENDENT VARIABLE RESPONSE SOURCE OF SUM OF SQUARES MEAN SQUARE F VALUE PROB > F R-SQUARE C.V. REGRESSION 9 ___78*186.29008683 [ 67.131.810009*5 ...... 51.1110* 0.0001 ______0 .6 1 9 1 U * 0 _ 11.2*102 X ' ERROR 2B3 *82*45.675783*6 1704.75503810 STP DEV RESPONSE MEAN CORRECTED TOTAL 292 1266631.96587031 _ ...... *1.26867930 367.30375 SOURCE DF SEQUENTIAL SS F VALUE PROB > F PARTIAL SS F VALUE PROB > F AG C 1 13*8*5.46106261 79.09961 0 .0 0 0 1 136096.79671387 79.8336* O.OCOl SUBJECTtAGE) 8 6*93*0.82902*22 ‘47.612*7 ' 0 .0 0 0 1 6*93*0.82902*22 *7.612*7 o .c c e i 186 Table 35 _IHRFSHOUP FOR VISIBILITY - REGRESSION ANALYSIS ANALYSIS OF VARIANCE TABLE , REGRESSION COEFFICIENTS , AND STATISTICS OF FIT FOR DEPENDENT VARIABLE CONTRAST SOURCE DF SUN OF SQUARES MEAN SQUARE F VALUE PROB > F R-SOUARS C.V. "EGRESSION______20 8_J.?3?U?62_. *.29660563 75.9*623. D. 0201_____ 0*8*46*626 ______31. 8 ei79_T ERROR 279 15.6050*390 0.0566*890 STD DFV CONTRAST MFAN CORRECTED TOTAL 299__ .101.73715652 ___ 0.23PU1030 . 0.7*65* SOURCE CF SEQUENTIAL SS F VALUE PROB > F PARTIAL SS F VALUE PRCE > F AGE 1. 0.636*1762 11.23**2 _ 0.0009 .. 0.636*1762 11.23**2 __ 0.0009 5U3JFCT SOURCE 8 VALUES T FOR HOlB-O PROB > 1T| STD ERR B STD B VALUES INTFRCEPT 0.755*6167 53.90000 0.0001 ______0.01*02*89. ______0.0 c 0,71*05000 36.7*333 0.0001 0.019*23*6 1.00116077 C2 -0.37213000 -19.1*893 0.0001 0.019*33*6 -0.52l75e92 Table 36 ______PDA FOB TWO PAYS FOR SO? ANALYSIS OP VARIANCE TABLE t REGRESSION COEFFICIENTS , AND STATISTICS OF FIT FOR DEPENDENT VARIABLE POA SOURCE OF SUM OF SQUARES MEAN SOUARE F VALUE PR OB > F R-SOUARE C.V. REGRESSION______31_ 32*.16902580 10.*5771051 1.0*015 C.*605 JL»325?Z.87Q_ _r.3,S0163._V ERROR 26 281.5128*75* 10.05*03027 STO DEV PDA MFAN CORRECTED TOTAL 59 605.70187333 3.170B0°C9 90.55233 " SOURCE OF SEQUENTIAL SS F VALUE PR08 > F PARTIAL SS F VALUE PROS > F SPEED 1 2.*3521*56 0.2*221 0.626* 0.117012*6 0. C116*_ 0.91*9 TRIAL ‘ 29 320.*9*53297 1.09921 0 .*070' 31«.73?prS75 1.C*661 0.A0A5 PAY 1 1.25*27827 0.12525 0.7261 1.25927827 0.12525 0.7261 SOURCE ______8_ VALUES______T_FORH0»P«0 ______PROB> |T |______$TP Epp...B _STT)..B.VALUES.-. INTERCEPT 90.82578*00 35.37*09 0.0001 2.56757*06 0.0 SPEED -0.0855317* -0.10788 0.91*9 0.79283228 -0.02029781 188 0.00 0.25 0.75 0.50 Time In Seconds 1.00 Ot'IOMO oiooino oelTs efr ne- ujc SOI Subject - ance Perform Task Foveal -D-CJOt-O JlO t-lDt-OC o n o o i o n i o n i Display On Time in Secondsin OnTime Display iue 56 Figure Standard Deviation of Deviation Standard Mean Response MeanResponse Total Errors Total Response Time Response Time 10 189 Number of Errors in 20 Trials 0 0.25 Time in Seconds 0.50 0.75 1.00 1.25 . 00 - t W 4 C O > 0 - t 0 4 C O I - H C C*| H ?H O M H O N O M ©*J H otooinomoinooinoin H H H Foveal Task Performance - Subject S02Subject - Performance Task Foveal Display On Time in Secondsin On Time Display Figure 57 Figure Standard Deviation,Standard of Mean Response Time MeanResponse Total Errors Total Response Time Response 190 Number of Errors in 20 Trials Time In Seconds 1.00 1*25 0* 50 0« 75 0.25 O O £ to O M o t Foveal Task Performance - Subject S04Subject - Performance Task Foveal © o Display On Time in Seconds in Time On Display iue 58 Figure * t to <0 o 4 to o t- to Mean Response MeanResponse Errors Total Standard Deviation of Deviation Standard o o Response Time Response Time CM IO to o - 191 10 0 Number of Errors in 20 Trials Time in Seconds 0.25 a 0.50 0.75 1*00 1*25 0 N o o t- 10 O W oelTs efr ne- Sbet SOS Subject - ance Perform Task Foveal o o Display On Time in Seconds in Time On Display iue 59 Figure t- in u> t- o o W 10 Standard Deviation of Deviation Standard Mean Response MeanResponse oa Errors Total to o Response Time Response Time t- o o 192 10 Number of Errors In 20 Trials Time In Seconds 0.25 0.75 0.50 1.00 1.25 0 S o n i ' O o t O A i o i o n i o o i rH Foveal Task Perform ance - Subject S06 Subject - ance Perform Task Foveal Display On Time in Seconds in Time On Display iue 60 Figure o d rH rH « Standard Deviation of Deviation Standard Mean Response Total Errors Total rH Response Time Response Time rH N 10 193 Number of Errors in 20 Trials Time in Seconds 0.50 0.25 0.75 1.00 1.25 ci ci o t Cl © o e-to to o Foveal Task Perform ance - Subject S07 Subject - ance Perform Task Foveal Clto Display On Time in Seconds in Time On Display iue 61 Figure o © Cl to Standard Deviation ofDeviation Standard Mean Response t- to Response Time Response Time © o Clto 10 194 Number of Errors In 20 Trials Time In Seconds 0.25 0.50 0.75 1.00 1.25 II) Ift U)O 1/3 to O o o O © #•••■• *•••• Cl ■ O Cn) *-4 »H •HfHrHMCg tH U) MN Foveal Task Perform ance - Subject SOS Subject - ance Perform Task Foveal ipa nTm n Seconds In Time On Display Figure 62 Figure to Standard Deviation ofDeviation Standard Total Errors Total Mean Response Response Time Response Time N O 0 1 195 Number of Errors in 20 Trials Time in Seconds 0.75 1.25 1 . 00 IO c*to to rH Foveal Task Perform ance - Subject S09 Subject - ance Perform Task Foveal H Display On Time In Seconds In Time On Display iue 63 Figure *-t * r-i o o r-i Total Errors Total Mean Response Time to 196 Number of Errors in 20 Trials Time In Seconds 0.25 0.50 0.75 1.00 1.25 1.50 04to M o o t-IO to o Foveal Task Perform ance - Subject S10 Subject - ance Perform Task Foveal l C IO Display On Time in Seconds in Time On Display iue 64 Figure o o I* to o o | C IO Standard Deviation of Deviation Standard Mean Response Response Mean Total Errors Total toO Response Time Response Time t-to o © Clto 10 197 Number of Errors in 20 Trials Residual >7.00000000 > - 00000000 0 0 0 0 0 0 .0 7 4.00000000 00000000 0 0 0 0 0 0 .0 8 0 0 0 0 0 5.000 1.00000000 Figure 65 Plot of Residual vs. PDA Predicted for S02 using Day 1 and Day 2 Data on PDA on Data 2 Day and 1 Day using S02 for Predicted PDA vs. Residual of Plot 65 Figure 45000 8*0000 65000 9.0000 75000 94.5000tli00 97.50000000 90.50000000 86.50000000 86*50000000 84.50000000 • •• • PDA - Predicted - PDA • •• • A A « ______A_ A A (Pilot Experiment, Stage 1) Stage Experiment, (Pilot A A H ------__A_A A~ A AA AA9 A ______A AA .A. ______A A A A Residual -4.00000000 -1.30000000 >6.50000000 6.oonooooo 3.30000000 1.00000000 ______ Figure 66 plot of Residual Versus PDA Predicted for Subject SOI Subject for Predicted PDA Versus Residual of plot 66 Figure a A. _ 23000 9.000C 47000 . 590C0 9.0000 98.30000000 97.10000000 95.9000C000 . 94.70000000 93.5000000C 92.30000000 . . A A A ‘ ______...... g- -g j A A A A A A A AAA A A A A - j — ~ ______A ______' A ' ' . A A A A A ------A (Pilot Experiment, Stage 2) Stage Experiment, (Pilot t *,. AAA A A A PDA - Predicted Predicted - PDA A A r ------A A , - — -— . A A ______------A A A A A A A ; ______A A . A ' ------_____ . * A . A A AB. ...... AA ______A A ------______ PDA UNDER NO LOAD AMO LOAO, 501 ANALYSIS OF VARIANCE TA3LE * REGRESSION COEFFICIENTS » AND STATISTICS OF FIT fCr DEPENDENT VARtABLE PDA SOURCE OF SUM OF SCUARES MEAN SQUARE F VALUF PROS > F R-SQUARE c.v. RFGRESSION______3l_ 202.18325534 6.52204049 ■.1.0*72,0 0.4532 0.53692240 -2*59263-? ERROR 28 174.36927800 6.22747421 STO OEV POA MEAN CORRECTED TOTAL 59 .376.55253333 ~y.**o549078~ 96.25333 SOURCE OF SEQUENTIAL SS F VALUF PROS > F PARTIAL SS F VALUE PROP > F SOEEO 1 29.51445592 4.739 39 0.03R1 l.C174!5?4 0,16338 _ 0.6891 TRIAL 29 ~ 171.00674015 '0.94690 '0.5581' '172.64 607491 G.9559e 0.5481 DAY 1 1.66205626 0.26689 0.6095 1.66205626 0.76669 0.6095 SOU»CE ______B_VALUES______T F0R_H0:8«0______PROB > IT!______STO FRR. B STO B..VALUES INTFRCF°T 95.38780523 44.04985 0.0001 2.16445106 0.0 SPFEO 0.27363142 0.40420 0.6891 0.67697525 0.0866517B 200 Table 38 PILOT EXPT.. STACE III. SUBJECT SOS ANALYSIS OF VARIANCE. FOR VARIABLE..PDA MEAN 87.A300000 . C.V. 1.9A758163? SOURCE OF SUM OF SOUARES MEAN SOUARE F_lOAD______.______2______A4.72A760 ______22.3623RO VIS 2 273.277960 136.638980 F_LOAO*VIS __ 21.7O38H0 5.AA8A70 RESIDUAL 36 10A.379AOO 2.E99A28 COROfCTEOJOTAL AA.______AAA, 176000 ______10,0“A TESTS —SOURCE ' ' - DFSUM OF SQUARES' MEAN SOUARE ~ F VALUE ...... PROB > F NUHEPATORt F_LOAD 2 AA.72A760 22.362380 7.7126* 0.0020 DENOMINATOR*‘RESIDUAL------36' 10A.379A00 2.899A28 ...... NUMERATORS__ VIS ______2 273.277960 136.638980 A7.12619 ... 0.0001 DENOMINATORS RESIDUAL 36 10A.37OA00 2.899A28 "NUMERATORS----F^LOAD*VTS------T ?K793efte 5.AA8A70..... 1.87915------'6.13A5 DENOMINATORS RESIDUAL 36 10A.379A00 2.899A28 201 • • 2 . 6 C O O O O O O ■Residual . 0 0 0 0 0 0 0 0.2 00000000 3 > 1.60000000 1.20000000 .00000000 Figure 67 Plot of Residual vs. PDA Predicted (Pilot Experiment, Stage HI) Stage Experiment, (Pilot Predicted PDA vs. Residual of Plot 67 Figure 15000 6.0000 55000 6.0000 89.50000000 67.50000000 85.50000000 63.50000000 81.50000000 PA Predicted - PDA _ "A" A _ _A A A ______ 91.30000000 202 Table 39 PILOT EXPERIMENT, STAGE lilt S03 ANALYSIS OF VARIANCE TABLE * REGRESSION COEFFICIENTS » AND STATISTICS OF FIT FOR DEPENDENT VARIABLE PDAT SOURCE OF SUM OF SQUARES MEAN SQUARE F VALUE PROS > F R-SCUARE C.V. REGRESSION______8______O.I2611635 0.01576454 14.31673 0 .0 0 0 1 0.76063112 2.65927 * ERROR 36 0.03964039 0.00110113 STD OEV PDAT MEAN CORPECTEO TOTAL 44 0.16575694 "0703318324" — 1.36931 SOURCE OF SEQUFNTIAL SS F VALUE PROS > F PARTIAL SS F VALUE PPCB > F LOAD 2 0.01576198 7.15720 0.0028 0.01576198 _ 7.15720 C.0C2R CONTRAST ----- ~ 2 — 0.10299028 "46.76583" "0.0001" "0.10299 0 28 " 46.76583" " C.CPC1 LCAD»CONTRAST 4 0.00736409 1.67194 0.1770 0.00736409 1.67196 0.1770 203 “ 2 -0.05000000 0 » § 0.10C00000 0.07000000 0.04000000 01000000 0 0 0 0 0 1 .0 0 02000000 0 0 0 0 0 2 .0 0 Residual) " ' A ' A" ' A A A A Figure 68 Residual vs. PDA (Transformed) Predicted (Pilot Experiment, Stage HI) Stage Experiment, (Pilot Predicted (Transformed) PDA vs. Residual 68 Figure .7000 1.30600000 1.27600000 A A A ...... ...... ...... ' "A TT" A A A A A A A A PDA Transform ed (Predicted) (Predicted) ed Transform PDA 1.3AOOOOCO A A 0 A A A A A ■ A 1.37200000 A A A A A 1.<.0400000 C A A 1 .42600000 A A A A A 204 Table 40 EFFECT OF SLIDE LOAD CONTRAST ON • PDA .* IN THF LAS. IS03I ANALYSIS OF VARIANCE FOR VARIABLE . KEAN 82.70X56?? C.V. 6.09572915 T SOURCE OF SUN OF SOUARES MEAN SQUARE SLIDE ...... 3 ..308.56386 102.85*62 LOAD 1 27.5*675 27.5*675 CONTRAST 1 1090.32825 1090.3 78 25 SLIDF*LOAD 3 9.76*18 3.2*806 SLIDE‘CONTRAST . 3 153.80503 51.268 3* LOAD*CONTRAST I 0.6*808 0 . 6*888 S L1DF *LOAD*CONTR AST 3 9.65P01 3.28600 ERROR 19 193.*31*6 10.18060 RESIDUAL ...... 16 183.573*5 11.673 3* CORRECTED TOTAL 31 17F3.868*2 57.56*1* TESTS SOURCE OF SUN OF SQUARES MEAN SQUARE F VALUF PROS > F NUMF»ATOR* SLIDE ' ...... 3 308.06386 102.85*62 10.10300 0.0005 DENOMINATORSERROR 19 193.631*6 10,16060 NUMERATORS LOAD 1 27.5*675 27.5*6 75 2.70581 0.1131 DENOMINATORS ERROR .... 19 193.631*6 10.18060 NUNFRATORt CONTRAST 1 1090.32825 1090.32825 107.09859 0 .0 0 0 1 DENOMINATORS ERROR 19 193.631*6 10.18060 numerator : SLIDE*LOAD _____ 3 9.76*18 3.2*806 0.31906 0.8132 DENOMINATORS ERROR 19 193.631*6 10.18060 NUN FRA TORS SLIOF*CONTRAST ...... 3 153.80503 51.26836 5.035R8 0 . 0 0 9 9 DENOMINATORS ERROR 19 193.631*6 10.16060 NUNFRATORt LOAO*CONTRAST ...... 1 0 . 6*888 0 . 6 * 8 8 8 0.06609 0.6301 DENOMINATORS ERROR 19 193.631*6 10.18060 -2.60000000 -*.40000000 6C00000 .60C A ?.«ocooooc 1.00000000 ,.80000000 £ (Residual 57000 7.5000 57000 8.5000 57000 90.75000000 85.75000000 80.75000000 75.75000000 70.75000000 65.75000000 Figure 69 Plot of Residual vs. PDA Predicted (Pilot Experiment, Stage 4) Stage Experiment, (Pilot Predicted PDA vs. Residual of Plot 69 Figure i A A A A A A A A PDA Predicted Predicted PDA \ A A A A A A A A A A A A A A A * A A A A A A A . ' ' o O) to Table 41 PI LOT EXPERIMENT. STAGE IV. SOS ANALYSIS OF VARIANCE TABLE f REGRESSION COEFFICIENTS , ANO STATISTICS OF FIT FOR OERENOENT VARIABLE POAT SOURCE OF SUN OF SQUARES MEAN SOUARE F VALUE PROB > F R-SOUARE c.v. REGRESSION 15 0.37763433 0.0251756? 8.17219 0 .0 0 0 2 ■SUM* 5* 5.43. _J,9Q719.*. ERROR 16 0.04929032 0.00308064 STD DEV PDAT MEAN CORRECTED TOTAL 31 0.42692465 0.05550356 I.4J055 SOURCE DF SECUFNT1AL SS F VALUE PROB > F PARTIAL SS F VALUE PROS > F SLIDE 3 0.P64B3517 7.0153? 0.0035 0.06483517 7.0153? C.OCp35 LOAD 1 " “ 0.G0P21020 2.66509 - " 0 .1 2 2 1 0 .0 0 621( 20 2.66509 0.1721 CONIRAST 1 0.274PC525 89.2P3P1 0 .0 0 0 1 0 .2748 0525 89.203F1 0.0001 SLIDE*LOAD 3 D.001P5604 0.2G0P3 0.8941 0.00185604 G.2CC63 C.fi°M SL IDF*CONTRAST 3 0.02507179 2.712R3 0.07RB 0.02E07179 2.712F3 G.07F f luao *ccntrast ...... ■ . r " 1 C.00129473 --- 0.42028 ---- 0.5Z6O 0.00129473 0 .42028“ “ “0.5760 StIPE*LOAD*CPNTRAST 3 0.00156115 0.16892 0.9153 0.00156115 0.16892 O.M53 207 PILOT EXPERIMENT, STAGE IV, S03 PLOT OF RES2 VS PDAT_H O.CPPOOOOO 0.05000000 A A A 0.02000000 A A A A RES2 A A A 0.01000000 A A A A A A A A A C.04000000 A A . A A 0.07000000 A - ••••:•• -s. 1.2P00C000 1.36000000 1.44000000 ■ 1.52000000 ' ' ^ 1.60000000 . PDA (Transformed) Predicted 208 Figure 70 Plot of Residuals vs. PDA (Transformed) predicted, Pilot Expt.,Stage IV Table 42 PERIPHERAL. DETECTION ANCLE EXPT. 1 ANALYSIS (lE VARIANCE FOR VARIABLE. PDJU -JJEAN- 92.7265927 C.V. 2.525*256? t- SCURCE OF SUH OF SQUARES MEAN SOUARE SUBJECT ______—I______LJA.02A22_____ 11A.07A216 ------ REPLICAT I 1.65*60 1.*5*599 LOAD ...... 2. — 32.27225 - 16.686672 CONTRAST 3 1219.6*222 *06.*80739 SUBJECT*LCAD_------_2_ 2.7-. 51359- —13.9 56797- SUE JE CT *CONl AAST 3 10.1*271 3.3809 05 LOAD*C(j’NTRA5T - . ______- 6 - 56.79151 - - 9,131913.. SUBJECT*Lf \r-«CONTRAST 6 11*.57«E6 19.0966** RFSIOUAL ’ . ------71_ 752.20929.. —10.7*9*25- CORRECTED TOTAL 95 23*0.9E195 2*.6*1915 TESTS SOURCE OF SUM OF SQUARES MEAN SOUARE F VALUE PROB > F NUMFRATPRI SUBJECT 1 116.07*22 116.07*216 ”l 0.798 i7~ oTo'cTo" cencminatpri RESIDUAL 71 763.20989 10.7*9*35 MI**FRATPR1 REPLICAT 1 1.*5*60 1.*5*599 0.13532 0.7151 C.ENCVINA'TCRt RESIDU AL . . . 71 _26i.20.9B9— 10.7*9*35. NUMERATOR* IPAD 33.37335 16 .666673 1.19559 0.6556 rfsr^lSATCRj SU9JCCT*LCA0 27.913 59 13.956797 NUMERATORS CONTRAST . __ -1219.6*222. *06»*E07 39.______120.22839- -C^GCil*- rENC^INATPRl SUBJtCT»CCNTRAST 10.1*271 3.3B0905 NUMERATORS SUFJFCT*LOAP 2 27.91359 13.956797 1.29837 0.2787 DENOMINATORS RESIDUAL 71 763.20989 10.7*9*35 M IM E* » TORS SU9JFCT*C0NT»AST .. _3. — 10. ^T I —3.360905- 0.31*52 0.8169 P E N C -IN * TORS RESIDUAL 71 TOS.209H9 10.7*9*35 v NUMERATORS LOAD*CONTKAST " 6 56.79151 9.131918 0.8*953 0.537* t'NCMINATCRS RESIDUAL 71 763.20989 10.7*9*35 209 NUMERATORS SUB JFCT *LCAP*CONTPAST 6 11*.57986 19.0966** 1.77653 0.1157 r e n p m n a i o r * RESIDUAL ______-71— J76 3. 209 B9— .10.7*9*35 Table 43 PfRIPHERAL DETECTION a n c le EXPT. 1 ANALYSIS OF VARIANCE TABLE , REGRESSION COEFFICIENTS t AND STATISTICS OF PIT FOR DEPENDENT VARIABLE PDA SCUPCE OF SUM OF SQUARES MEAN SOUARE F VALUE PROB > F R-SOUARE C.V. REGRESSION______25_ 1598.34592445 63.93383699 6.02633 0 .0 0 0 1 0.68276730 3.51227 * ERROR 70 7A2.63602271 10.60906604 STD DEV PDA MEAN CORRECTED TOTAL 95 2340.98196716 J.25715920^ 92.7365° SOURCE DF SEQUENTIAL SS F VALUE PROB > F PARTIAL SS F VALUE PRC8 > F SPFED 1 154.5°60C,144 14.57704 0.0003 20.57JP6620 1.o?027 0.16P2 SUejECT ------! ~U2.«55821°4' 10.64708- 0.0017- —114.19S3422T' -10.81143- 0.CC16- 0eflICat 1 7.03127558 0.66276 0.4183 3.273662C2 0.30E57 0.5803 LOAD 2 9.75083855 0.4SO55 0.6392 19.°°Cvi?707 0.54258 0 . 6 6 ?? CONTRAST 3 1124.0194,8183 35.34453 0.0001 1129.70018210 35.497P5 0.CC01 SURJEC T*LOAD------Z' 10.77029752- —0i5076fT 0.6098- 10i5R70770?- -0.40873- -0.61«1 SUEJFCT»CONTRAST 3 9.12008496 0.28658 0.8365 9.09 220279 0.78567 O.f 371 LOAD*CONT°AST 6 50.61177662 0.79510 0.5784 50.1741518e 0.7PP73 0.je36 SU3JECT*L0AD*C0NTRAS 6 116.58944600 1.86302 0.0989 118.56944600 1.86302 0.0989 SOURCE t VALUES T FOR H0lB«0 PROB > (TI STD ERR 8 STD B VALUES INTERCEPT 95.04508328 42.46938 0 .0 0 0 1 2.25680460 0.0 SPEED -1.01802953 --1.39258 0.1682 0.73104044 ------0.11C49075 210 10.00000000 - - 0.00000000 1 6.0C00GG00 2.00000000 00000000 0 0 0 0 0 0 .0 2 00000000 0 0 0 0 0 0 .0 6 Residual B£.aOflOCtOCQ ______------Figure Figure A A. A -A k J Bft.lOflOOQOQ A ------A. 1 7 Residual vs. PDA Predicted (Experiment 1) (Experiment Predicted PDA vs. Residual ______A A PERIPHERAL DETECTION ANGLE EXPT, f ANALYSIS OF VARIANCE FOR VARIABLE POAT . MEAN ... 1.22033*5* - C.V., .. 6.3829650* \ ... SOURCE OF SUM OF SQUARES MEAN SQUARE SUBJECT 1 ft.0«31372ft ft -ftfi3137279 RE»L!CAT 1 0.00200**7 0.00200**69 LOAD ----- 2 -- .. - 0.01918*2* 0.009592118 ------CONTRAST 3 0.68323886 0.2277*62 86 S1IBJFCT*! OAfl ...... ? C,01?3A79 3« SUBJCCT*CONTRAST 3 0.00861*25 0.002871*18 LOAO*CONTRAST ____ 6 ..... - 0.03628608 -0.0060*76 80 _ .... ------SUBjrCT»LOAD*CCNTRAST 6 0.07ft 6*623 0.0131I603e RFN irVJAL , . 71 . ft.*7n7RR?1 . 0.006067**0 . CORRECTED TOTAL 95 1.366725*9 0.0'*3B65fl* TESTS SOU®CF OF SU« OF SQUARES MEAN SOUARE r VALUE PROS > F SU“EPATCRJ SUBJECT ...... ‘ ' T 0.08313728 6 .08 3 *372 79 ~ 13.70220 0.000.7 T E*lCvI*IATOPt RESIDUAL 71 0.*3078R21 0.006067**0 rjL«E®ATP»J REPLICAT 1 0.00200**7 0.(0200**69 0.33036 0.574? OENC»INATCRt RESIDUAL 71 0.A307HR?1_._ 0.006047**0 NUMERATOR* LOAD 2 0.01916*2* 0.009592118 0.77*31 0.5636 CE\'CM1NATCP» SUBJECT*LOAD 2 0.02*775ftft 0.012387939 VUMEPATPP.S CONTRAST . 0.68323SR6 0.2277*6286 79.31*07 O.ftftR* r'NOPlVATORs SUBJE CT *CCNTRAST 3 0.00861*25 0.002871*18 NUMERATORS SUBJECT*LOAO ' — - 2 0.02*77588 0.012387939 2.0*171 0^1353 DENCrlNATCRt RESIDUAL 71 0.A307RP21 0.006067**0 NUMERATORS SL'EJFCT*CONTRAST . . .. 3 ------0.00861*25 0.002871*18 ... 0.67325 0.706C rENCMlf.ATCRS RFSIDUAL 71 0.*3078821 0.006067**0 NUMERATORS LCAO*CON7RAST 6 0.03628608 0.0060*7680 0.9967* 0.56*9 DENOMINATORS RFSIC'UAL 71 0.*3078821 0.006067**0 212 NUMERATORS suejfct *load *contrast 6 0.07869623 0.013116038 2.16171 0.0563 Qi.NCKiNATOP.S RESIDUAL...... -71 0.43078821——0^006067440 ---- Table 45 ...... PERIPHERAL DETECTION ANGLE EXPT. 1 ANALYSIS OF VARIANCE 1 ABLE ' REGRESSION COEFFICIENTS t AND STATISTICS OF F I T FOR DEPENDENT VARIABLE POAT SOURCE OF SUM OF SQUARES KEAN SOUARE F VALUE PROB > F R-SQUARE C.V. REGRESSION______25 0.95429369 0 .0 3 9 1 7 1 7 5 6 .4 7 8 7 0 0.0001 0 .6 9 8 2 3 3 6 1 6.29996 T ERROR 70 0.41243181 0 .0 0 5 S 9 1 8 B STO DEV POAT KEAN CORRECTED TOTAL 95 1 .3 6 6 7 2 5 4 9 0.07675860* ‘ " 1 .2 2 0 3 3 SOURCE OF SEQUENTIAL SS F VALUE PROB > F PARTIAL SS F VALUE PRC9 > F SPEED 1 0.11159178 18.53922 0.0001 0 . 0 1 835640 3.11554 0.0819 SUBJECT 1 O.OEOP^SIB 13.72094 0 .C 0 0 4 " 0.08704043 13.97471" C.0CC4 REPLICAT 1 C."0700461 1.1 9996 0 .2 7 9 3 0.CO 393755 0.66745 0.4167 LOAD 2 0.00934006 0.3 6 H 31 0 .6 4 8 6 0 .01C0699 5 0.65456 0.5669 CONTRAST 3 0.61452495 35.04962 0.0001 0.67466603 35.340*4 0.0001 SUE JECT*LDAD ...... ?- - 0.00RR3229 -0.74Q53- 0.5194- ‘0.00875555" “ 0.763 06- 0.5166. SUSJECT*CONTRAST 3 G.00758460 0.42°C9 0.7365 O.OC756752 0.47795 0.7376 LOAP*C GNTR AST 6 0.03319383 0.93997 0 .5 2 6 0 0.03796133 0 .9 9 2 9 4 0.5215 SUBJECT*LOAO*CONTRAS 6 0.08133667 2.30081 0 .0 4 3 3 0.08133667 2.20081 0.0432 SOURCE 8 VALUES T FOR M 0:B »0 PROB > | T | STO ERR B STD 8 VALUES INTFRCEPT 1.12748384 21.19963 0.0001 0.05318413 O.C SPEED — 0.03040858 1.76309 ------— 0 .0 8 1 9 ------0.01727779 ------0.13659037 213 0.22000000 A A A C.UOOOOOO A A A A A A A A A A A 0.06000000 A A ■—< A A d A A A A 3 A AA A *3 A A •H A A B A A A A A «n A A A A A A ■ A A -o.iecoocoo 1.06000000 1.16000000 1.2*000000 1.32000C00 . 1.*0000000 1.40000000 PDA (Transformed) Predicted Figure 72 Residual vs. PDA (Transformed) Predicted (Experiment 1) 214 Table 46 PERIPHERAL DETECTION ANC.LE. EXPERIMENT 2 - ...... AOE«Y ------...... ANALYSIS OF VARIANCE FORVARIABLE PDA — MEAN 93.1641728 C.V. 2.63900630 X SOURCE OF SUM OF SQUARES MEAN SQUARE 2 RA>.An?A?___ 281.201310- LOAD 2 29.49066 14.7493 30 ...... CONTRAST ______...... 2 ----1. 387.76221 193.883104 SESSION 2 10.51469 5.25°R47 PERIOD ...... ______. 2 ..------6.87772- _____ 3 .42 68 59 . ------. ------LOAO-*CfiNTRAS1 A 7.8«J5A2 1.973854 ______RESIDUAL ...... 66 ____ 308.95385 6.044755 ...... ---- CORPECTED TOTAL 80 1603.91016 17.5488 77 TESTS SnuRCE OF SUM OF SQUARES MEAN SQUARE F VALUE PROB > F NUMERATOR* SUPJ ' ...... 2 ' 562.4026? 261.201310 46.51988 0 .C001 OENCM INATGRi RESIDUAL 66 29e.95385 6.044755 NUMERATOR s LOAD 2 29.49866 14.749330 2.44002 C.0931 CEUCYINATOR* RESIDUAL ...... - ______66 ... — 398.95385 6.044755 NUMERATOR: CONTRAST 2 387.76221 193.881104 22.07427 0.0 0 0 1 DENOMINATORS RESIDUAL ...... ' 66 *98.95385 6.044755 *;U“EPATOftI SESSION ...... 2. ____ 10.51969 5.259847 ...... 0.87015 0.5733 DENOMINATORS RESIDUAL 66 398.95385 6.044755 NL“E°ATCR: . PERIOD ...... 2 6.87772 3.43 8* 59 ...... 0.56890 0.5741 DENOMINATORS RESIDUAL 66 398.95385 6.044755 NUMERATORS LCADRCONTRAST . . ______...6 ______7.69542 1.973854. .. -. . 0.326 54 - 0.8547 CFNCMIHATO* I RESIDUAL • 66 398.95385 6.0447 55 Table 47 PERIPHERAL DETECTION ANGLE, EXPERIMENT 2 ACE»A ANALYSIS TF VARIANCE FOP VARIA8LE PDA . . L . MEAN 8 9 .4 6 1 1 8 5 2 C .V . ... 4 .0 0 1 5 4 5 7 2 X SOURCE OF SUM OF SQUARES MEAN SOUARF .SUM - - ____ - ______2 _ 320.10626 .... 1 6 0 .0 5 3 1 4 1 ------LOAD 2 155.33675 77.668374 CONTRAST . 2 609.27677. 304.638385 ... SESSION ? 16.8379? 7.418460 pSRIQD ~ - ______2-- 1.96113 0.96056.5 ...... LCAD»CONTRAST 4 66.21716 16.554290 EESIOl'AL ...... -...... 66 645.80218 . „ 12.815184...... CORRECTED TOTAL 60 2013.53619 25.169227 TESTS SOURCE ' OF SUN OF SOUARFS MEAN SQUARE F VALUE PROB > F NUMERATOR* SUM """ 2 320.10428 160.053141 ~ 12.4R9J*, o .o e ri DENOMINATOR* RESIDUAL 66 845.80218 12.R151P4 NUMERATOR* LOAO 2 155.334,75 77.646374 6.06065 0.0042 DENOMINATOR* RESIDUAL . . 66 .... 645.80216 12.P151B4 numerator * CONTRAST 2 609.27677 304.638365 23.77167 0.0CC1 DENOMINATOR* RESIDUAL"" 66 ...... """ 845.80716 "" 12.815184 NUMERATOR* SESSION ...... 2 .. . - 14.83792 7.418960 0.57802 0.5685 r ENOMINATOR* RESIDUAL 66 645.80218 12.815184 "numepatcr * — PERIOD 2 1.96113 ' 0.9*0565" ’ 0.07652 0.4360 DENOMINATOR* RESIDUAL 66 645.80218 12.815184 NUvr°ATCR S LOAD*CCNT«AST .4 . 66.21716 16.554290.. 1.29177 0.2815 DENOMINATOR* RESIDUAL ■ 66 845.80218 12.815184 217 Table 48 PERIPHERAL DETECTION ANGLE. EXPERIMENT 2 ANALYSIS OF VARIANCE FOR VARIABLE PDA______MEAN____ .91.3126790 C.V...... 3.31900*19 t , SOURCE OF SUM OF SQUARES MEAN SOUARE- —AGE------1------555.3*076------555.3*0762 ------SUB J! AGE I * SB?. 50890 220.627226 LOAD ...... — 2 ------27.*3622 13.718109 ------CONTRAST 2 98*.57962 *92.289808 LOAD*CONTRAST ... - .. . ------.___— *, ------.... 52.55715 - 13.139786 . ... — ...... AGERLOAD 2 157.39919 78.699596 AGE*CONTRAST ...... ------______2------12.*5936 - - 6.229681 ...... ------AGE»LUAO*CONTRAST * 21.555*3 5.388858 ..SESSION ...... - 9 ------22.*3500. . -,.11.217501 ____ . ------PERIOD 2 7.361*8 3.680739 . RESIDUAL ..... _.. 136 ____ -12*9.15600 . . 9.18*971 ------CORRECTED TOTAL 161 3972.78911 2*.675709 — ------. . . ------...... —• ------...... TESTS SOURCE OF SUMOF SQUARES MEAN SQUARE F VALUE PROB > F NUMERATORS AGE ~ I 555.3*076 555.3*0762 2.51710 0.1871 DENOMINATOR: SU8JI AGE) 882.50890 220.6272 26 * ______- NUMERATOR s SUBJlAGE I * 882.50890 220.627226 2*. 020*6 G.0 CG1 CENCM lN'ATCR: RESIDUAL 136 ______12*9.15600 ...... 9.18*9 71 - ...... - - - • NIIMERATCR! LOAO 2 27.*3622 13.718109 1.A935* 0.2267 DENOMINATOR! RESIDUAL 136 12*9.15600 9.18*971 lUJMERATCK: CONTRAST _ 2 ------.98*.57962 -*92.289808 . .. 53.59732 . 0 .0 0 0 ! DE\f“I\AnK! RESIDUAL 136 12*9.15600 9.18*971 NUMERATOR! LOAD*CONTRAST _. .* ------— 52.55715 ------13.139286 1.63052 ------0.2261 DENOMINATOR! RESIDUAL 136 12*9.15600 9.18*971 NUMERATOR! AGFALOAD 2 157.39919 '78.699596' 8.56830 0.0006 rSNC«INATOR! RESIDUAL 136 12*9.15600 9.18*971 NUMERATOR! Af.EACCNTRAST 2 12.*5936 6.229681 0.67825 0.5T38 DENOMINATOR! .RESIDUAL. . 116 12*9.15600 _____ 9.18*971 . , , . . -----. — . . . NUMERATOR! ACE»L0AD*C0NTRAS1 * 21.555*3 5.388858 0.58670 0.6761 DENOMINATOR! RESIDUAL 136 12*9.15600 9.18*9 71 NUMERATOR! SESSION _ __ 2______22.*3500 ____ 11.217501 .... . - 1.22129 ------. 0.2977 OFNCM INATtiRs RESIDUAL 136 12*9.15600 9.18*971 NUMERATOR! PERIOD 2 7.361*8 3.660739 0.60073 0.676* TENO-'INA IORs RESIDUAL 136 12*9.15600 9.18*971 ... Table 49 ______PERIPHERAL DETECTION. ANCLE•--EXPERIMENT S------ANALYSIS OP VARIANCE TABLE • REGRESSION COEFFICIENTS • AND STATISTICS OP PIT FOR DEPENDENT VARIABLE PDA SOURCE OP SUM OF SQUARES MEAN SOUARE F VALUE PROB > F R-SQUARF C.V. REGRESSION______2 6 _____ .... 2781.19516250. -106.96904471- -12.11891—. --O.0001 - -0.7000M11 —3.25362 *- ERROR 135 1191.59395081 8.82662186 STD DEV PDA KEAN -CORRECTED TOTAL ______161 ... 3972.78911331 2.97096312 91.31268 SOURCE DF SEQUENTIAL SS F VALUE PROB > F PARTIAL SS F VALUE PROS > F SPEEO 1 . - 565.48491783 . . - .64,06584 0 .0 0 0 1 57.56705296 6.*?l4l C .0118 AGE 1 354.01256497 40,67284 O.OCOi 430.04 54 3 2 7 3 48.72141 C.CCC1 SUBJ{AGE) 4 619.67897064 17.5814? 0 .0 0 0 1 617.72727363 17. 49600 C.OOOl LOAD 2 45 .5 5 87PP06 2 .58243 0.0774 40.P6395759 2.314F1 0.1006 CONTRAST ____ 2 . ___ -1000.23267821 ______56.65999. ___ __ 0 .0 0 0 1 . . . 995.95239C58 - - 56.41753 ----- C.CCC1 5FSSI0N 2 13.53262628 0.76h59 0.5293 15.?56019Ct G. f.6470 0.573? PERIOD 2 2.95016855 0.1671? 0.8472 0.4.’ 751511 0.0242? 0.9768 LOAD *CONTR AST 4 41.97703586 1.16893 0.3181 44.331919(0 1.25563 0.2899 AGE*LOAD ______2 ...... - „ 99.86995665 ...... 5.65732 . . ... 0.0047 99.82944646 - - 5.656(2 0.0C47 AGE*CONTRAST 2 11.52868618 0.65306 0.5268 11.52 297790 C.65274 G.5270 AGE*LOAD*CONTRAST 4 21.33902328 0.60439 0.6634 21.33902328 0.6C439 C.6634 SOURCE ...... a VALUES T FOR H0!B«0.-...... PROB > |Tj . STO ERR B - - STO b VALUES INTFRCEPT 93.93835254 89.0« 643 0.0001 1.05434474 0.0 SPEED -0.95615539 -2.55371 0.0118 0.37441876 -0.15464471 - -lO.CCOOOOOO. - - .0000000 0 0 0 0 0 0.00 1 00000000 0 0 0 0 0 0 .0 2 00000000 0 0 0 0 0 0 .0 6 00000000 0 0 0 0 0 0 .0 6 .0 0 0 0 0 0 2.000 Residual ~ ~ *...... ‘ ~~ ~ ------~ A ' ' A ~ A A 14000 8.0000 B6000 9.0000 58000 99.60000000 95.80000000 92.20000000 8B.60000000 85.00000000 81.40000000 PA A A A * A A ' Figure 73 Plot of Residual vs. PDA Predicted (Experiment 2) (Experiment Predicted PDA vs. Residual of Plot 73 Figure A A A A ~ 'A r - ~ A ----- A A ...... A B A A ' ------A A A A A A A A A A A ~ A ' ‘ A ‘ “ ' ‘A' ~ A A A A — A- A A A A - A A — A A A A B AA A A AA A BA AA^^A A AAA A A A A A A A AA A A A A eA A AA A S A A A A A A A D Predicted PDA AA A AA ...... -B ...... AA A ...... • - - • - A B A A A " ~ " " A AAAA B A-- .A ... - B A ...... B A A A AAA A AA AA A A A AAA A A A AA A A a A ‘ V A A V ~ " •■ ; •■“ " "" ~ A - ...... -B - - A A A----- A A A * - - ...... A ------• Blank Page * BIBLIOGRAPHY Abemethy, C. N. and Leibowitz, H. W.,"The Effect of Feedback on Luminance Thresholds for Peripherally Presented Stimuli," Perception and Psycho physics, Vol. 10, No. 3, Sept. 1971, pp. 172-174. Allen, M. J ., Vision and Highway Safety, Chilton Book Co., Philadelphia, Pa., 1970. Alpern, M., "Movement of the Eyes, " (in) Davson, The Eye, Vol. 3, Academic P re ss, New York, 1962, pp. 3-229. Alpern,. M. ,"Rod Vision, " Chapter 3 (in) The Assessment of Visual Function, edited by A. M. Potts, The C. V. Mosby Co., St. Louis, Mo., 1972. Antes, J. R. and Edwards, D. C., "Information Processing in the Visual Periphery, Bulletin of Psychonomic Science, Vol. 1, No. 5B, 1973, pp. 351-353. Babington Smith, B ., "An Unexpected Effect of Attention in Peripheral Vision, Nature, Vol. 189, No. 4766, March 4, 1961, pp. 776. Babington Smith, B. /'Effect of Attention in Peripheral Vision." Nature, Vol. 191, No. 4789, August 12, 1961, pp. 732-733. Bartlett, N. R., Bartz, A. E ., and Wait, J. V., "Recognition Time for Symbols in Peripheral Vision, Highway Research Bulletin 330, 1962, pp. 87-91. Bartz, A. E ., "Eye-Movement Latency, Duration, and Response Time as a Function of Angular Displacement," J. of Experimental Psychology, Vol.64, No. 3, 1962, pp. 318-324. Bartz, A. E. ,"Eye and Head Movements in Peripheral Vision: Nature of Com pensatory Eye Movements, " Science, Vol. 152, June 17, 1966, pp. 1644-1645. Bartz, A. E ., "Fixation Errors in Eye Movements to Peripheral Stimuli," J. of Experimental Psychology, Vol. 75, No. 4, Dec. 1967, pp. 444-446. Behaine, J. O. and Wick, Jr., R. L., "Field of View of Aircraft Pilots Wearing Eye Glasses, " Working Paper, Department of Preventive Medicine, The Ohio State University, 1976. 221 222 Belt, B. L., Driver Eye Movements as a Function of Low Blood Alcohol Con centrations, Technical Report, Driving Research Laboratory, Department of Industrial and Systems Engineering, The Ohio State University, 1969. Bhise, V. D. and Rockwell, T. H., Development of a Methodology for Evalua tion Visual Clarity Information Previews in Driving, Paper presented at the 40th National Meeting of ORSA, Anaheim, CA., Oct. 1971. Bhise, V. D. and Rockwell, T. H., Role of Peripheral Vision and Time Sharing in Driving, Paper presented at the American Association for Automotive Medicine, Colorado Springs, CO., Oct. 1971. Bhise, V. D., "Visibility Distance Prediction in Rain, Environmental and Safety Engineering, Ford Motor Company, 1976. Bloomfield, J. R. and Howarth, C. L, "Testing Visual Search Theory, " (in) Leibowitz, H. W. fed.), Image Evaluation, Proceedings of NATO Symposium, 1969, pp. 203-214. Bloomfield, J. R./'Peripheral Acuity with Complex Stimuli at Two Viewing Distances," AGARD Conference Proceedings, No. 100, A ir to Ground Target Acquisition, Paper B6, Brussels, Belgium, 1972. Bloomfield, J, R ., "Visual Search in Complex Fields: Size Differences between Target Disc and Surrounding Disc, " Human Factors, Vol. 14, No. 2, April 1972, pp. 139-148. Bloomfield, J, R ,, "Experiments in Visual Search, fin) Visual Search, National Academy of Sciences, 1973. Bloomfield, J. R., Theoretical Approaches to Visual Search, Paper presented at the International Symposium on Human Factors in Quality Control in Buffalo, New York, 1974. Bloomfield, J. R ., Classifying Studies of Visual Inspection, Paper presented at the International Symposium on Human Factors in Quality Control in Buffalo, New York, 1974. Bloomfield, J. R., Marmurek, H. H., and Traub, B. G., "Color and Texture Differences in Embedded Target Visual Search Situations, " Proceedings of the 18th Annual Meeting of the Human Factors Society, Oct. 1974, pp. 158-170. Borkenhagen, J. M. /'Rotary Acceleration of a Subject Inhibits Choice Reaction Time to Motion in Peripheral Vision, J. of Experimental Psychology, Vol. 102, No. 3, March 1974, pp. 484-487. 223 Bouman, M. A., "On Foveal and Peripheral Interaction in Binocular Vision, " Optica Acta, Vol. 1, No. 4, Feb. 1955, pp. 177-183. Brown, B ., "Dynamic Visual Acuity, Eye Movements, and Peripheral Acuity for Moving Targets," Vision Research, Vol. 12, 1972, pp. 305-321. Burg, A., "Lateral Visual Field as Belated to Age and Sex, " J. of Applied Psychology, Vol. 52, No. 1, 1968, pp. 10-15. Bursill, A. E., "The Bestriction of Peripheral Vision during Exposure to Hot and Humid Conditions, " Quar. J. of Experimental Psychology, Vol. 10, Part 3, 1958, pp. 113-129. Carney, L. G., "Effect of Hypoxia on Central and Peripheral Corneal Thick ness and Corneal Topography," The Australian J. of Optometry, Vol. 58, Feb. 1975, pp. 61-65. Colenbrander, M. C., "Visual Acuity, Visual Field and Physical Ability, " Opthalmologica, Vol. 171, 1975, pp. 100-108. Connolly, P. L ., "Visual Considerations of Man, the Vehicle, and the Highway, " Highway Vehicle Safety, SAE PT 13, 1968. Cornsweet, D. M. ,"Use of Cues in the Visual Periphery under Conditions of Arousal, " J. of Experimental Psychology, Vol. 80, No. 1, 1969, pp. 14-18. Cornsweet, T. N., Visual Perception, Academic Press, 1970. Corso, J. F., Hie Experimental Psychology of Sensory Behavior, Holt, Rine hart and Winston, Inc., 1967. Cotton, R. J. and Nutt, J. A., "The Perception of Movements of Reflectorized Targets under Conditions of Low Illumination," The British J. of Physio logical Optics, Vol. 19, No. 1, Jan. -M arch 1962. Cumming, R. W. and Croft, P. G.,"Human Information Processing under Vary ing Task Demands," Ergonomics, Vol. 16, No. 5, 1973. Also in Man under Stress, edited by A. T. Welford, Taylor and Francis Ltd. London, 1974. Dale, H. C. A., "Factors Affecting the Choice of Strategy in Searching," Ergonomics, Vol. 7, No. 1, Jan. 1964, pp. 73-82. Danielson, R. W., "The Relationship of Fields of Vision to Safety in Driving, " Amer. J. of Ophthalmology, Vol. 44, No. 5, Part I, 1957. 224 Davidoff, J. B ., Differences in Visual Perception; The Individual Eye, Academic Press, 1975. Devlin, W. A. and Roe, R. W., "The Eye Ellipse and Considerations in the Driver's Forward Field of View,"SAE Paper 680105, Automotive Engi neering Congress, Detroit, Mich., Jan. 8-12, 1968. Dirken, J. M. (ed.), Functional Age of Industrial Workers, Netherlands Institute for Preventive Medicine TNO, Wolters-Noordhoff Publishing, Groningen, 1972. Ditchburn, R. W., Eye-Movements and Visual Perception, Clarendon Press, Oxford, London, 1973. Drury, C. G., "The Human Factor in Industrial Inspection, " Quality Progress, Dec. 1974, pp. 14-19, 32, 33. Edwards, D. C. and Goolkasian, P. A., "Peripheral Vision Location and Kinds of Complex Processing, " J. of Experimental Psychology, Vol. 102, No. 2, 1974, pp. 244-249. Ellerbrock, V. J ., "A Comparison of Peripheral Stereoscopic and Visual Acuities,''American J. of Optometry and Arch, of American Academy of Optometry, Vol. 26, No. 12, Dec. 1949, pp. 530-537. Ellis, K., "Methods of Scanning a Large Visual Display, " Occupational Psychol ogy, Vol. 42, 1968, pp. 181-188. Engel, F. L. "Visibility, Conspicuousness and Attention, " Abstract in Ophthal mologies, Vol. 171, 1975, p. 41. See Vision Research, Vol. 11, 1971, pp. 563-576, Enoch, J. M., "Effect of the Size of a Complex Display upon Visual Search, " J. of the Optical Society of America, Vol. 49, No. 3, March 1959, pp. 280-286. Erikson, C. W. /'Location of Objects in a Visual Display as a Function of the Number of Dimensions on Which the Objects Differ, " J. of Experimental Psychology, Vol. 44, No. 2, August 1952, pp. 56-60. Erikson, C. W,, "Object Location in a Complex Perceptual Field," J. of Experi mental Psychology, Vol. 45, No. 3, March 1953, pp. 126-132. Erikson, C. W., "Partitioning and Saturation of Visual Displays and Efficiency of Visual Search, " J. of Applied Psychology, Vol. 39, No. 2, April 1955, pp. 73-77. 225 Erickson, R. A ., "Relation between Visual Search Time and Peripheral Visual Acuity, " Human Factors, Vol. 6, April 1964, pp. 165-177. Ernest, J. T. and Krill, A. E. ,"The Effect of Hypoxia on Visual Function: Psychophysical Studies, " Investigative Ophthalmology, Vol. 10, No. 5, May 1971, pp. 323-328. Feinberg, R., "A Study of Some Aspects of Peripheral Visual Acuity: Part I, " Amer. J. of Optometry and Arch, of American Academy of Optometry, Vol. 26, No. 2, Feb. 1949, pp. 49-56. Feinberg, R., "A Study of Some Aspects of Peripheral Visual Acuity: Parts II, in, and IV," Amer. J. of Optometry and Arch, of American Academy of Optometry, V ol., No. 3, Feb. 1949, pp. 105-119. Forbes, L. M., "Geometric Vision Requirements in the Driving Task, " 1970 International Automobile Safety Conference Compendium, published by Society of Automotive Engineers, 1970, pp. 677-689. Ford Motor Company, "Field of View from Automotive Vehicles, SAE Publica tion, SP-381, 1973. Fry, G., Private Communication, 1976. Gasson, A. P. and Peters, G. S., "The Effect of Concentration upon the Ap parent Size of the Visual Field in Binocular Vision, Part I, " The Optician, Vol. 148, No. 3848, Jan. 1, 1965, pp. 660-663. Part II, The Optician, Vol. 149, No. 3849, Jan. 8, 1965, pp. 5-12. Gibson, J. J ., The Senses Considered as Perceptual Systems, Houghton Mifflin Company, 1966. Goolkasian, P. and Bunt, A. A. ,"Foveal and Peripheral Vision in a Dynamic Stimulus Situation, " Institute for Perception TNO, Report Number IZF 1973-C13, 1973. Gould, J. D. and Schaffer, A., "Eye-Movement Patterns during Visual Informa tion Processing, " Psychonomic Science, Vol. 3, 1965, pp. 317-318. Gould, J. D. and Schaffer, A., "Eye-Movement Patterns in Scanning Numeric Displays, " Perceptual and Motor Skills, Vol. 20, 1965, pp. 521-535. Gould, J. D. and Schaffer, A., "The Effects of Divided Attention on Visual Monitoring of Multi-channel Displays, " Human Factors, Vol. 9, No. 3, 1967, pp. 191-202. 226 Gould, J. D., "Eye Movements during Visual Search, " IBM Research Report, RC 2680, July 1969. Gould, J. D. and Peeples, D. R. ,"Eye Movements during Visual Search and Discrimination of Meaningless, Symbol, and Object Patterns," J. of Exper imental Psychology, Vol. 85, No. 1, 1970, pp. 51-55. Gould, J. D. ,"Eye Movements during Visual Search and Memory Search," J. of Experimental Psychology, Vol. 98, No. 1, April 1973, pp. 184-195. Gould, J. D. and Carn R., "Visual Search, Complex Backgrounds, Mental Counters and Eye Movements, " Perception and Psychophysics, Vol. 14, No. 1, August 1973, pp. 125-132. Gould, J. D., "Looking at Pictures," IBM Research Report, RC 4991, August 1974. Graham, C. H. (ed,), Vision and Visual Perception, John Wiley and Sones, 1965. Graham, C. H. ."Movement and Space Perception, " (in)' Visual Science, edited by J. R. Pierce and J. R. Levene, Indiana Univ. Press, 1971, pp. 179- 196. Grindley, G. C. and Townsend, V., "Voluntary Attention in Peripheral Vision and its Effects on Acuity and Differential Thresholds, " Quar. J. of Experimental Psychology, Vol. 20, Part I, February 1968. Haber, R. N., "Visual Information Storage, " Paper presented at the National Research Council Committee on Vision Annual Meeting in May 1970. Published in Visual Search, National Academy of Sciences, Washington, D.C., 1973. Haber, R, and Hershenson, M., The Psychology of Visual Perception, Holt, Rinehart and Winston, Inc., 1973. Haines, R. F ., "Peripheral Visual Response Time and Retinal Luminance- Area Relations, " Amer. J. of Optometry and Physiological Optics, Vol. 52, Feb. 1975, pp. 85-95. Harper, N., "Driving and Wearing Spectacles, " (Letter), British Medical J., Vol. 3, p. 1073, 1966. Harrington, D. O., The Visual Fields; A Textbook and Atlas of Clinical Perimetry, Third Edition, The C. V. Mosby Co., St. Louis, 1971. 227 Hay, J. C., ’’Visual Adaptation to an Altered Correlation between Eye Move ment and Head Movement, " Science, Vol. 160, April 26, 1968, pp. 429-430. Held, H. H., "Peripheral Acuity, " British J. of Physiological Optics, Vol. 16, No. 3, July 1959, pp. 126-143. Hennessy, R. T. and Leibowitz, H, W., "The Effect of a Peripheral Stimulus on Accommodation, " Perception and Psychophysics, Vol. 10, No. 3, Sept. 1971, pp. 129-132. Heron, A. and Chown, S., Age and Function, Little, Brown & Co., Boston, 1967. Highway Safety Research Center, University of North Carolina, "Are Narrow Visual Fields Related to Accident Involvement?" Highway Safety High lights, 1975. Highway Safety Research Center, University of North Carolina, "Looking at the Problem of Tunnel Vision, " The Accident Reporter, 1974. Howarih, C. L and Bloomfield, J. R., "Towards a Theory of Visual Search, " AGARD Conference Proceedings, No. 41, London Paper A2, 1968. Howarth, C. L and Bloomfield, J. R., "A Rational Equation for Predicting Search Times in Simple Inspection Tasks," Psychonomic Science, Vol. 17, No. 4, 1969, pp. 225-226. Howarth, C. L and Bloomfield, J. R., "Search and Selective Attention, " British Medical Bulletin, Vol. 27, No. 3, Special Issue on Cognitive Psychology, Sept. 1971, pp. 253-258. Human Factors Research Inc., Training and Evaluation of the Peripheral Vision of Motor Vehicle Drivers, Research sponsored by NHTSA (in progress), California, 1976. Huntley, M. S., "Effects of Alcohol and Fixation-Task Difficulty on Choice Re action Time to Extrafoveal Stimulation, ’’ Quar. J. of Studies on Alcohol, Vol. 34, 1973, pp. 89-103. Institute of Traffic Engineers, Traffic Engineering Handbook, edited by J. E. Baerwald, Third Edition, 1965. Jay, B. S., "The Effective Pupillary Area at Varying Perimetric Angles, " Vision Research, 1_, 1961, pp. 418-424. 228 Johnson, C. A, and Leibowitz, H. W., "Practice, Refractive Error, and Feedback as Factors Influencing Peripheral Motion Thresholds, " Perception and Psychophysics, Vol. 15, No. 2, 1974, pp. 276-280. Johnston, D. M., "Search Performance as a Function of Peripheral Acuity, " Human Factors, Vol. 7, No. 6, Dec. 1965, pp. 527-535. Kahwaji, M. N., Mateossian, R., and Griffin, J. R., "Visual Vield Extent Re lated to Test Target Size and Illuminance, Amer. J. of Optometry, Vol. 50, No. 1, Jan. 1973, pp. 40-45. Kaluger, N. A. and Smith, Jr., G. L., "Driver Eye-Movement Patterns under Conditions of Prolonged Driving and Sleep Deprivation, " (in) Night Visi bility and Driver Behavior, Highway Research Record, No. 336, Dec. 1970. Kantowitz, B. H., "Double Stimulation, " Chapter 3 (in) Human Information Pro cessing; Tutorials in Performance and Cognition, edited by B. H. Kantowitz, Lawrence Erlbaum Associates, Publishers, Hillsadale, N.J., 1974. Keeney, A. H., "Assessment of Special Visual Function, " Chapter 9 (in) Hie Assessment of Visual Function, edited by A. M. Potts, The C. V. Mosby Co., St. Louis, 1972. Kephart, N. C. and Chandler, R. E ., "Changes of the Visual Field in a Pursuit Tracking Task," The Optometric Weekly, March 22, 1956, pp. 507-509. Kerr, J. L., "Visual Resolution in the Periphery, " Perception and Psycho physics, Vol. 9, No. 3B, March 1971, pp. 375-378. Kite, C. R. and King, J. N., "A Survey of the Factors Limiting the Visual Fields of Motor Vehicle Drivers in Relation to Minimum Visual Field and Visibility Standards," British J. of Physiological Optics, Vol. 18, No. 2, April 1961, pp. 85-107. Knowles, W. B., "Operator Loading Tasks, " Human Factors, Vol. 5, No. 2, April 1963, pp. 155-161. Kobrick, J. L., "Effects of Physical Location of Visual Stimuli on Intentional Response Time, " J. of Engineering Psychology, Vol. 4, No. 1, 1965, pp. 1-8. Kobrick, J. L. and Dusek, E. R., "Effects- of Hypoxia on Voluntary Response Time to Peripherally Located Visual Stimuli, " J. of Applied Physiology, Vol. 29, No, 4, Oct. 1970, pp. 444-448. 229 Kochhar, D. S. and Fraser, T. M., "The Effect of a Simulated Driving Task on Signal Monitoring in the Peripheral Visual Field, " Paper presented at the Annual Meeting of the Ergonomics Research Society, Cardiff, Wales, April 11-14, 1972. Kochhar, D. S. and Fraser, T. M., "Peripheral Visual Performance in a Simu lated Driving Task... Some Quantitative Aspects, " Paper presented at the First International Conference on Driver Behavior, Zurich, Switzer land, October 8-12, 1973. Kochhar, D. S., Human Peripheral Vision: Empirical Models Defining Opera tional Attributes, Ph. D. Thesis, Dept, of Systems Design, University of Waterloo, Ontario, July 1974. Kretovics, T ., Analysis of Frequency and Amplitude of Head Movements in Driver Mirror Sampling, M. S. Thesis (in preparation), The Ohio State •University, 1976. LeGrand, Y., Light, Color, and Vision, translation of LeGrand (1948) by R. W. Hunt, J. W. Walsh, and F. R. Hunt, Chapman & Hall, London, 1957. LeGrand, Y., Form and Space Vision, Indiana University Press, Bloomington, Ind., 1967. Leibowitz, H. W., Visual Perception, Ihe Macmillan Company, 1965. Leibowitz, H. W., Johnson, C. A., and Isabelle, E., "Peripheral Motion De tection and Refractive Error," Science, Vol. 177, Spet. 29, 1972, pp. 1207-1208. Leibowitz, H. W., "Detection of Peripheral Stimuli under Psychological and Physiological Stress, " Paper presented at the Annual Meeting of the National Research Council Committee on Vision in May 1970. Published (in) Visual Search, National Academy of Sciences, Washington, D. C., 1973. Low, F. N., "Studies on Peripheral Visual Acuity," Science, Vol. 97, No. 2530, July 25, 1943, pp. 586-587. Low, F. N., "The Peripheral Visual Acuity of 100 Subjects, " Amer. J. of Physiology, Vol. 140, No, 1, October 1943, pp. 83-88. Low, F. N., "The Peripheral Visual Acuity of 100 Subjects under Scotopic Conditions, " Am er. J. of Physiology, Vol. 146, No. 1. April 1946, pp. 21-25. 230 Low, F. N., "Some Characteristics of Peripheral Visual Performance, " Amer. J. of Physiology, Vol. 146, No. 4, July 1946, pp. 573-584. Low, F. N., "The Development of Peripheral Visual Acuity during the Process of Dark Adaptation, " Amer. J. of Physiology, Vol. 146, No, 4, July 1946, pp. 622-629. Low, F. N., "The Peripheral Motion Acuity of 50 Subjects, " Amer. J. of Physiology, Vol. 148, No. 1, Jan. 1947, pp. 124-133. Low, F. N., "Peripheral Visual Acuity, "A.M. A Arch, of Ophthalmology, Vol. 45, No. 1, Jan. 1951, pp. 80-99. Mackworth, N. H., "Visual Noise Causes Tunnel Vision, " Psychonomic Science, Vol. 3, pp. 67-68, 1965. McColgin, F. H., "Movement Thresholds in Peripheral Vision, " J. of the Optical Society of America, Vol. 50, No. 8, Aug. 1960, pp. 774-779. Moray, N., Attention; Selective Processes in Vision and Hearing, Academic Press, 1970. Moses, R. A., Alder's Physiology of the Eye; Clinical Application, 5th edition, Chapters 18-21, The C. V. Mosby Co., St. Louis, Mo., 1970. Moskowitz, H., Sharma, S., and McGlothlin, W., "Effect of Marihuana upon Peripheral Vision as a Function of the Information Processing Demands in Central Vision, " Perceptual and Motor Skills, Vol. 35, 1972, pp. 875-882. Moskowitz, H., Sharma, S., and Schapero, M., "A Comparison of the Effects of Marijuana and Alcohol on Visual Functions, " (in) Current Research in Marijuana, edited by Mark F. Lewis, Academic Press, 1972. Moskowitz, H., and Sharma, S., "Effects of Alcohol on Peripheral Vision as a Function of Attention, " Human Factors, Vol. 16, No. 2, A pril 1974, pp. 174-180. Mourant, R. R. and Rockwell, T. H., "Mapping Eye-Movement Patterns to the Visual Scene in Driving; An Exploratory Study, " Human Factors, Vol. 12, No. 1, 1970, pp. 81-87. Mourant, R. R. and Rockwell, T. H., "Strategies of Visual Search by Novice and Experienced Drivers," Human Factors, Vol. 14, Issue 4, 1972, pp. 325-335. 231 Murch, G. M., Visual and Auditory Perception, The Bobbs-Merrill Co., Inc., New York, 1973. National Academy of Sciences, Visual Search, collection of papers presented in a Symposium conducted at The Spring Meeting, 1970. Committee on Vision, Division of Behavioral Sciences, National Research Council, Washington, D. C ., 1973. National Research Council, Committee on Vision, First Interprofessional Standard for Visual Field Testing, National Academy of Sciences, Washington, D. C., 1975. Neisser, U. ."Visual Search, M Scientific American, Vol. 210, No. 6, June 1964, pp. 94-102. Newman, M. /'V isual Field, " Chapter 2 (in) The A ssessm ent of Visual Function, .edited by Albert M. Potts, The C. V. Mosby Co., St. Louis, Mo., 1972. Mewsome, L. R., "Visual Angle and Apparent Size of Objects in Peripheral Vision, " Perception and Psychophysics, Vol. 12, No. 3, Sept. 1972, pp. 300-304. Pirenne, M. H., Vision and the Eye, Science Paperbacks and Chapman and Hall, Ltd., 1967. Potts, A. M., (ed.), The Assessment of Visual Function, The C. V. Mosby Co., St. Louis, 1972. Reeves, F. B. and Bergum, B. O., "Perceptual Narrowing as a Function of Peripheral Cue Relevance, " Perceptual and Motor Skills. Vol. 35, 1972, pp. 719-724. Rentschler, L , Hilz, R ., and Grimm, W., "Processing of Positional Informa tion in the Human Visual System, " Nature, Vol. 253, No. 5491, Feb. 6, 1975, pp. 444-445. Richards, O. W., "Visual Needs and Possibilities for Night Automobile Driving," American Optical Corporation, Southbridge, M ass., Aug. 1967. Riggs, L. A., "Visual Acuity, " (in) Vision and Visual Perception, edited by C. H. Graham, et al., John Wiley & Sons, Inc., New York, 1965. Roberts, H. J ., The Causes, Ecology and Prevention of Traffic Accidents, Charles C. Thomas, Publisher, Springfield, 111., 1971. 232 Robinson, G. H., "Visual Search by Automobile Drivers, " paper presented to Symposium on Psychological Aspects of Driver Behavior, Vol. 1, held at Noordwykerhout, The Netherlands, Aug. 2-6, 1971. Robinson, G. H. and Subelman, H., Visual Search for a Peripheral Digital Dis play during a Central Manual Control Task, Tech. Report TR-75-6, Department of Industrial Engineering, Univ. of Wisconsin, Madison, Wis. March 1975. Robinson, G. H. and Bond, J. R., Visual Dynamics when Interrupting a Control Task to Search for a Peripheral Target, Tech. Report TR-75-21, Dept, of Industrial Engineering, Univ. of Wisconsin, Madison, Oct, 15, 1975. Robinson, G. H., Koth, B. W., and Ringenbach, J. P ., Dynamics of the Eye and Head during an Element of Visual Search, Tech. Report TR-75-5, Dept, of Industrial Engineering, Univ. of Wisconsin, Madison, Oct. 1975. Rockwell, T. H., "Eye Movement Analysis of Visual Information Acquisition in Driving: An Overview, " Australian Road Research Board Proceedings, P ap er No. 948, Vol. 6, No. 3, 1972, pp. 316-331. Rockwell, T. H. and Balasubramanian, K. N., "Carbon Monoxide Effects on Highway Driving Performance: An Investigation of the Effects of Twelve Percent COHb on the Nighttime Performance of Young and Aged Drivers, " Paper presented to American Association for Automotive Medicine, Nineteenth Annual Conference, San Diego, California, 1975. Rockwell, T. H., Bhise, V. D., and Nemeth, Z. A., Development of a Com puter Based Tool for Evaluating Visual Field Requirements of Vehicles in Merging and Intersection Situations, Final Report, RF 3306, Systems Research Group, Department of Industrial and Systems Engineering, The Ohio State University, 1973. Rockwell, T. H., et al., Utility of Peripheral Vision to Motor Vehicle Drivers, (report in progress), Project RF 4224A1, Systems Research Group, Dept, of Industrial and Systems Engineering, The Ohio State University, 1976. Rockwell, T. H., Wick, R. L., and Balasubramanian, K. N., Effects of Noxious Gases on Driver Performance, Final report submitted by the Systems Research Group, Dept, of Industrial and Systems Engineering, The Ohio State University to Dept, of Transportation, National Highway Traffic Safety Administration, August 1974. Rogers, J. G., "Peripheral Contrast Thresholds for Moving Images, " Human Factors, Vol. 14, No. 3, June 1972, pp. 199-205. 233 Rolfe, J. M., "Hie Secondary Task as a Measure of Mental Load, " (in) Measurement of Man at Work, edited by W, T. Singleton, J. G. Fox, and D, Whitfield, Taylor and Francis Ltd., London, 1972. Ronchi, L. and Molesini, G., "Depth of Focus in Peripheral Vision, " Ophthalmic R esearch, Vol. 7, 1975, pp. 152-157. Rubin, M. L. and Walls, G. L., Fundamentals of Visual Science, Charles C. Thomas, publisher, Springfield, 111., 1969, Rubin, M. L., "Visual Acuity, " Chapter 1 (in) The Assessment of Visual Func tion, edited by Albert M. Potts, The C. V. Mosby Co., St. Louis, 1972. Sanders, A. F ., The Selective Process in the Functional Visual Field, Institute for Perception RVO-TNO, National Defence Research Organization TNO, Soesterberg, The Netherlands, 1963. Sanders, A. F ., "Some Aspects of the Selective Process in the Functional Visual Field, " Ergonomics, Vol. 13, No. 1, 1970, pp. 101-117. Saugstad, P. and Lie, L , "Training of Peripheral Visual Acuity, " Scandinavia^ J. of Psychology, Vol. 5, 1964, pp. 218-224. Schioldborg, P ., "Attention and Visual Identification Time, " Scandinavian J. of Psychology, Vol. 12, No. 4, 1971, pp. 289-294. Schmidt, L , "Visual Considerations of Man, the Vehicle, and the Highway, P a rt I, " (in) Highway Vehicle Safety, SAE PT 13, 1968. Scientific American, Perception; Mechanisms and Models, W. H. Freeman & Company, 1972. Senders, J. W., "On the Distribution of Attention in a Dynamic Environment, " Acta Psychologica, Vol. 27, 1967, pp. 349-354. Senders, J. W., "The Estimation of Operator Workload in Complex Systems, " Chapter 7 (in) Systems Psychology, edited by K. B. DeGreene, McGraw Hill Series in Management, 1970. Senders, J. W., "Visual Scanning Behavior, " (in) Visual Search, Symposium conducted at the Spring Meeting 1970, Committee on Vision, Division of Behavioral Sciences, National Research Council, National Academy of Sciences, Washington, D. C., 1973. 234 Smith, H. P, R. and Weale, R. A., "Obstruction of Vehicle-Dri vers' Vision by Spectacle Frames, " British Medical J., Vol. 3, p. 445. Smith, S. W. and Semmelroth, C., An Experimental Study Relating Search Time to Peripheral Visual Discrim inability of Objects, Report 2900-95-M, Institute of Science and Technology, The University of Michigan, Nov. 1961. Thomson, A., "Effect of Attention in Peripheral Vision, " Nature, Vol. 191, No. 4789, Aug. 12, 1961, p. 732. Versace, J ., "Driver Vision: State of the Art, " 1970 International Automobile Safety Conference Compendium, published by Society of Automotive Engi neers, 1970, pp. 617-628. Von Wright, J. M. and Mikkonen, V., "The Influence of Alcohol on the Detec tion of Light Signals in Different Parts of the Visual Field, " Scandinavian J. of Psychology, Vol. 11, 1970, pp. 167-175. Weale, R. A., "Problems of Peripheral Vision, " British J. of Ophthalmology, Vol. 40, No. 7, July 1956, pp. 392-415. Weale, R. A., The Aging Eye, Harper and Row, Publishers, London, 1963. Weltman, G. and Egstrom, G. H., "Perceptual Narrowing in Novice Divers, " Human Factors, Vol. 8, No. 6, Dec. 1966, pp. 499-506. Weymouth, F. W., "Visual Sensory Units and the Minimal Angle of Resolution, " A m er. J. of Ophthalmology, Vol. 46, No. 1, P a rt IL, July 1958, pp. 102-113. Williams, L. G., "Target Conspicuity and Visual Search, " Human Factors, Vol. 8, No. 1, Feb. 1966, pp. 80-92. Williams, L. G., A Study of Visual Search Using Eye Movement Recordings, Honeywell, Inc., St. Paid, Minn., Research Report 12009-IR1, Feb. 1966, Research Report 12009-IR2, 1967. Williams, L. G., "The Effects of Target Specification on Objects Fixated during Visual Search, " Acta Psychologica, Vol. 27, 1967, pp. 355-360. Williams, L. G., "Studies of Extrafoveal Discrimination and Detection, " Paper presented at the Annual Meeting of the National Research Council Committee on Vision, May 1970. Published in Visual Search, National Academy of Sciences, Washington, D. C., 1973. 235 Wolf, E ., Effects of Age on Peripheral Vision, Highway Research Board Bulletin 336, Night Visibility, 1962. Wolf, E ., "Studies on the Shrinkage of the Visual Field with Age, " Highway R esearch Record, No. 164, Night Visibility, 1967. Zahn, J. R. and Haines, R. F., "The Influence of Central Search Task Lumin ance upon Peripheral Visual Detection Time, " Psychonomic Science, Vol. 24, No. 6, Sept. 25, 1971, pp. 271-273. Zunzanyika, X. K. and Drury, C. G., "Effects of Information on Industrial Inspection Performance, " Paper presented in the International Symposium on Human Factors in Quality Control, held in Buffalo, N. Y., 1974.