PERFORMANCE IMPROVEMENT METHODS FOR TERRAIN DATABASE
INTEGRITY MONITORS AND TERRAIN REFERENCED NAVIGATION
A thesis presented to
the Faculty of the
Fritz J. and Dolores H. Russ
College of Engineering and Technology
of
Ohio University
In partial fulfillment
of the requirements for the degree
Master of Science
Ananth Kalyan Vadlamani
March 2004 This thesis entitled
PERFORMANCE IMPROVEMENT METHODS FOR TERRAIN DATABASE INTEGRITY
MONITORS AND TERRAIN REFERENCED NAVIGATION
BY
ANANTH KALYAN VADLAMANI
has been approved for
the School of Electrical Engineering and Computer Science
and the Russ College of Engineering and Technology by
Maarten Uijt de Haag
Assistant Professor of Electrical Engineering and Computer Science
R. Dennis Irwin
Dean, Russ College of Engineering and Technology VADLAMANI, ANANTH K. M.S. March 2004. Electrical Engineering and Computer Science
Performance Improvement Methods for Terrain Database Integrity Monitors and Terrain Referenced Navigation (115pp.)
Director of Thesis: Maarten Uijt de Haag
Terrain database integrity monitors and terrain-referenced navigation systems are based on performing a comparison between stored terrain elevations with data from airborne sensors like radar altimeters, inertial measurement units, GPS receivers etc. This thesis introduces the concept of a spatial terrain database integrity monitor and discusses methods to improve its performance. Furthermore, this thesis discusses an improvement of the terrain-referenced aircraft position estimation for aircraft navigation using only the information from downward-looking sensors and terrain databases, and not the information from the inertial measurement unit. Vertical and horizontal failures of the terrain database are characterized. Time and frequency domain techniques such as the Kalman filter, the autocorrelation function and spectral estimation are designed to evaluate the performance of the proposed integrity monitor and position estimator performance using flight test data from Eagle/Vail, CO, Juneau, AK, Asheville, NC and Albany, OH.
Approved: Maarten Uijt de Haag Assistant Professor of Electrical Engineering and Computer Science To my parents
ACKNOWLEDGEMENTS
During my study at Ohio University, I’ve worked on many projects, as part of the coursework under the guidance of my teachers, but none was as challenging or as extensive as this one: the Masters’ thesis. Challenging, because it led me to explore and understand concepts as an engineer and I wish to thank everyone who have helped me during the course of this research.
I express my sincere thanks to my advisor, Dr. Maarten Uijt de Haag, who got me involved in research and encouraged me at every stage. Thank you Maarten, for your patience, attention and the faith that motivated me to go on. Also, for the numerous reviews and inputs, in the past for various conference papers and now for this thesis, which have helped to make this work meaningful. We have brought many a project to fruition and I’m sure we will continue to do so in future.
I am grateful to my thesis committee members Dr. Michael Braasch, Dr. Frank van Graas and Dr. William (Gene) Kaufman for their time and effort in reviewing my thesis and their useful comments. I thank Dr. Braasch for the great learning experience during all the courses I took with him these last two and a half years that have contributed immensely to my understanding and my research. I thank Dr. van Graas for introducing and laying a solid foundation to the concepts that will remain with me throughout my career.
I thank Jacob Campbell for a lot of things: patiently explaining his thesis to me, providing me with the data and some initial routines to work on, for the useful discussions that helped me burst through a plateau phase in my research, for his support and for simply being there, so I knew I could run up to him in case of problems.
I thank Steve Young for his useful inputs during conferences and Dr. Robert Gray for his initial research that got it all started. I am thankful to the NASA B757 ARIES flight crew for their support and expertise during the EGE flight trials. I am deeply grateful to the Ohio University King-Air C90 pilots Brian Branham and Jamie Edwards for the flight tests at JNU, also to Dr. Richard McFarland for the flight tests at AVL and KUNI conducted on the Ohio University DC3 and the chief of airborne laboratories, Jay Clark, for his help and support during the said flight- testing. Support from the terrain database providers: NIMA, NGS, and Jeppesen is greatly appreciated. The research presented in this thesis was supported and funded through NASA under Cooperative Agreement NCC-1-3511.
I thank the faculty, staff and all my colleagues at the Avionics Engineering Center who have all been part of this learning experience. In fact, I believe with all our interactions, I have learned as much in the break-room, hallways and student offices at AEC, as I have in formal coursework. It’s a great place with great people. I thank all my friends who have supported, tolerated, and motivated me, who have heard me out, who have set examples in all aspects of life and from whom I learn constantly.
And most of all, I thank my parents, Ramanand and Kanaka Durga Vadlamani, for their love and encouragement and for making me the person I am. Although words cannot express my gratitude, I owe everything to them, my brother, Ravi, who has been there to share my joys and sorrows and my family who have constantly supported me. It is wonderful to be amongst you.
1 Opinions, interpretations, conclusions and recommendations are those of the author and are not necessarily endorsed by NASA 7
TABLE OF CONTENTS
Abstract...... 3 Acknowledgements ...... 5 List of Tables ...... 9 List of Figures...... 10 List of Acronyms...... 13 1 Introduction...... 15 1.1 Contributions ...... 16 1.2 Outline of the Thesis ...... 17 2 Synthetic Vision Systems ...... 18 2.1 SVS Component Description ...... 21 2.1.1 Sensors...... 22 2.1.2 Terrain Databases ...... 22 2.1.3 Displays ...... 23 2.2 SVS Predecessors ...... 24 2.2.1 Ground Proximity Warning Systems (GPWS) ...... 24 2.2.2 Terrain Awareness and Warning System (TAWS) ...... 24 3 Terrain Database Integrity Monitor ...... 26 3.1 Statistical Method...... 30 3.1.1 Formulation of Hypotheses...... 31 3.1.2 Test Statistic for Decision Making...... 33 3.1.3 Pseudo-Random Noise Analysis ...... 36 3.2 Vertical Domain Integrity Monitor...... 37 8 3.3 Horizontal Domain Integrity Monitor ...... 38 3.4 Spatial Integrity Monitor ...... 41 4 Terrain Referenced Navigation...... 43 5 Theoretical Background ...... 52 5.1 The Kalman Filter...... 55 5.1.1 Test Statistic Using Kalman Estimates ...... 58 5.1.2 Pseudo-Random Noise Analysis Revisited...... 60 5.2 Autocorrelation Function Estimation ...... 62 5.3 Modern Spectral Estimation ...... 65 5.3.1 Blackman – Tukey Spectral Estimation ...... 66 5.3.2 Maximum Entropy Spectral Estimation...... 66 6 Spatial Failure Detection and Estimation ...... 70 6.1 EGE Flight Test and System Description...... 71 6.1.1 NASA ARIES Triple Redundant Radar Altimeter ...... 72 6.1.2 Slant Range Errors ...... 72 6.1.3 Altitude Dependent Errors...... 74 6.1.4 Mid-Value Select...... 74 6.2 Terrain Database Integrity Monitor...... 75 6.2.1 Vertical Domain Integrity Monitor...... 76 6.2.2 Horizontal Domain Integrity Monitor ...... 78 6.2.3 Spatial Integrity Monitor ...... 81 6.3 Application to Terrain Referenced Navigation...... 82 6.3.1 Lateral Position Estimation...... 83 6.3.2 Vertical Bias Estimation ...... 88 6.3.3 Spatial Position Estimation ...... 89 7 Additional Flight Test Results...... 92 7.1 Flight Test Environment...... 93 7.1.1 JNU Flight Test...... 93 7.1.2 AVL Flight Test...... 93 7.1.3 KUNI Flight Test ...... 94 7.2 Terrain Database Integrity Monitor...... 94 7.2.1 Vertical Domain Integrity Monitor...... 95 7.2.2 Horizontal Domain Integrity Monitor ...... 95 7.2.3 Spatial Integrity Monitor ...... 95 7.3 Applications to Terrain Referenced Navigation...... 99 7.3.1 Lateral Position Estimation...... 99 7.3.2 Vertical Bias Estimation ...... 101 7.3.3 Spatial Position Estimation ...... 102 8 Summary and Conclusions...... 104 9 The Road Ahead...... 107 References...... 109 Appendix...... 113 A.1 Terrain Database Specifications ...... 113 A.2 Kalman Filter Design Parameters...... 114 9
LIST OF TABLES
Table 3.1 Decision Making in Hypothesis Testing ...... 32 Table 3.2 Over-Bounded Absolute Disparity Distributions for Sensors and Data...... 32 Table 4.1 A comparison of Several TRN Systems...... 51 Table 6.1 Mean E-N Coordinates of Horizontal Position Estimates...... 84 Table 6.2 Mean E-N Coordinates of Horizontal Position Estimates of Figure 6.13 ...... 85 Table 6.3 Mean E-N Coordinates of Horizontal Position Estimates of Figure 6.14 ...... 85 Table 6.4 Mean E-N Coordinates of Horizontal Position Estimates of Figure 6.15 ...... 87 Table 7.1 Mean E-N Coordinates of Horizontal Position Estimates of Figure 7.6 ...... 99 Table 7.2 Mean E-N Coordinates of Horizontal Position Estimates of Figure 7.7 ...... 100 Table 7.3 Mean E-N Coordinates of Horizontal Position Estimates of Figure 7.8 ...... 101 Table 7.4 Mean E-N Coordinates of Horizontal Position Estimates of Figure 7.10 ...... 103 Table A.1 Terrain Database Specifications...... 113
10
LIST OF FIGURES
Figure 2.1 – SVS Concept Architecture Overview...... 21 Figure 2.2 – Depiction of (a) Linear Error and (b) Circular Error in the DEM ...... 23 Figure 3.1 – Sample Synthetic Vision System Fault Tree ...... 29 Figure 3.2 – Measurement of Synthesized Terrain Profile ...... 30
Figure 3.3 – Distribution of T for H0 and H1...... 34 Figure 3.4 – Threshold value for chi-square statistic...... 35 Figure 3.5 – Non-Centrality parameter look-up plot ...... 35 Figure 3.6 – Operating Characteristics (OC) Curve ...... 36 Figure 3.7 – T Statistic values for Approaches to R/W 25 at EGE...... 37 Figure 3.8 – T value for 0, 25m and 35m biases...... 38 Figure 3.9 – Region of Missed Detection within –20 to 20 Resolution Grid in Latitude and Longitude ...... 40 Figure 3.10 – Space Envelope of Missed Detection ...... 41 Figure 3.11 – Sample Space Envelope Plot at EGE ...... 42 Figure 4.1 – The TERCOM System ...... 44 Figure 4.2 – Measurements and Terrain Correlation in the TERCOM System...... 44 Figure 4.3 – The SITAN System ...... 46 Figure 4.4 – Frequency Domain Terrain Correlation Scheme of [37]...... 48 11
Figure 5.1 – P and Q as a function of µB ...... 59
Figure 5.2 – Autocorrelation Sequence of Kalman Estimates for µB of 34.2m, 25m and 15m...... 60
Figure 5.3 – Correlation Coefficient (ρ) of Kalman Estimates for µB of 34.2m, 25m and 15m ....61
Figure 5.4 – OC Curve of Kalman estimates for µB of 34.2m, 25m and 15m ...... 61 Figure 5.5 – Autocorrelation Function of a WSS Random Process...... 62 Figure 5.6 – Autocorrelation of finite length constant signal ...... 63 Figure 5.7a, 5.7b – Autocorrelation of 50 sample length Gaussian noise having mean 15 and σ = 18.9...... 63
Figure 5.8 – Rxx(0) Estimation Using a Straight Line Fit ...... 64 Figure 5.9a, 5.9b – Mean and Standard Deviation of Bias Estimate ...... 65 Figure 5.10 – Spectral Estimates Using Blackman-Tukey and Maximum Entropy Methods ...... 67 Figure 5.11a – Zero Frequency Power Estimates for Various Vertical Biases...... 68 Figure 5.11b – Higher Frequency Power Estimates for Various Vertical Biases...... 68 Figure 6.1 – Runway 07 and 25 Approaches and Departures at EGE ...... 71 Figure 6.2 – Illustration of Plumb-Bob Height and Slant Range...... 73 Figure 6.3 – Illustration of Altitude Dependent Errors...... 74 Figure 6.4 – Mid-Value Select Scheme ...... 75 Figure 6.5 – Different Test Statistics During Approach to R/W 25 at EGE for Biases of 0, 25m and 35m...... 77 Figure 6.6 – Augmentation Scheme for DEM Integrity Monitor ...... 78 Figure 6.7 – Improvement in RoMD within –20 to 20 Resolution Grid Points in Latitude and Longitude ...... 79 Figure 6.8 – Improvement in RoMD at Locations shown in Figure 3.9...... 80 Figure 6.9 – RoMD during Entire Flight Path at EGE...... 81 Figure 6.10 – Improvement/Reduction in the Volume of Space Envelopes...... 82 Figure 6.11 – Terrain Navigation using Various Statistics and their E-N Error Comparison ...... 83 Figure 6.12 – Horizontal Position Fixes on the DEM...... 84 Figure 6.13 – Horizontal Position Fixes on the DEM with Horizontal Bias ...... 85 Figure 6.14 – Horizontal Position Fixes on the DEM with Horizontal and Vertical Biases ...... 85 Figure 6.15 – Horizontal Position Fixes on the DEM with Horizontal and Vertical Biases using Equations (6.1) and (6.2)...... 86 Figure 6.16 – Terrain Information Metric Variation...... 87 Figure 6.17 – East and North Direction Error Variation with ‘I’...... 88 12 Figure 6.18a – Vertical Bias Estimates using Mean and ACF Estimators...... 89 Figure 6.18b – Pitch and Roll Angles...... 89 Figure 6.19 – Illustration of Spatial Position Estimation...... 90 Figure 6.20 – Horizontal Position Fixes using Spatial Position Estimation ...... 91 Figure 6.21 – Spatial Position Estimation of East, North and Vertical Biases ...... 91 Figure 7.1(a) – Ohio University King Air C90 Flying laboratory, ...... 93 Figure 7.1(b) – DEM Integrity Monitor Experiment (DIME) equipment ...... 93 Figure 7.2 – Ohio University DC-3 Flying laboratory...... 94
Figure 7.3 – T and TKF values during flight segments at JNU, AVL and KUNI for biases of 0, 25m and 35m...... 96 Figure 7.4 – Improvement in RoMD at EGE, JNU, AVL and KUNI...... 97 Figure 7.5 – Reduction in Space Envelope using the Kalman Filter method ...... 98 Figure 7.6 – Horizontal Position fixes on the DEM at JNU ...... 99 Figure 7.7a,b and c – Horizontal Position fixes on the DEM at AVL without biases ...... 100 Figure 7.7d,e and f – In the presence of both Horizontal and Vertical biases ...... 100 Figure 7.8a,b and c – Horizontal Position fixes on the DEM at KUNI without biases ...... 101 Figure 7.8d,e and f – In the presence of both Horizontal and Vertical biases ...... 101 Figure 7.9 – Vertical Bias Estimates using Mean and ACF Estimators at JNU, AVL and KUNI ...... 102 Figure 7.10 – Horizontal Position Fixes on the DEM at JNU, AVL and KUNI ...... 102 Figure 7.11 – Spatial Estimation of Vertical Biases at JNU, AVL and KUNI ...... 103 Figure A.1 – Illustration of a shortest interval and its associated probability (area) under a normal curve...... 114 Figure A.2 – Steady state Kalman filter parameter values...... 115
13
LIST OF ACRONYMS
ADS-B Automatic Dependant Surveillance - Broadcast ACF Autocorrelation Function AD Absolute Disparity AGL Above Ground Level AHRS Attitude Heading Reference Set ATC Air Traffic Controller BT Blackman - Tukey CaB Commercial and Business CDTI Cockpit Display of Terrain Information CEP Circular Error Probability CFIT Controlled Flight Into Terrain COTS Commercial Off The Shelf DEM Digital Elevation Model DGPS Differential GPS DoD Department of Defense DTED Digital Terrain Elevation Data DWL Downward-Looking ECEF Earth Centered Earth Fixed EGPWS Enhanced Ground Proximity Warning Systems EKF Extended Kalman Filter EVS Enhanced Vision Systems FAA Federal Aviation Administration FHA Functional Hazard Assessment FLIR Forward Looking Infrared FLTA Forward Looking Terrain Avoidance 14 FTA Fault Tree Analysis FWL Forward-Looking GA General Aviation GNSS Global Navigation Satellite System GPS Global Positioning System GPWS Ground Proximity Warning System HDD Head Down Display HMD Head Mounted Display HUD Head Up Display ILS Instrument Landing System IMC Instrument Meteorological Conditions INS/IRS/IMU Inertial Navigation/Reference/Measurement System/Unit LAAS Local Area Augmentation System LEP Linear Error Probability LiDAR Light Detection And Ranging MDB Minimum Detectable Bias MESE Maximum Entropy Spectral Estimation MMWR Millimeter Wave Radar MPDLIM Multiple Path Downward Looking Integrity Monitor MSE Mean Squared Error MSL Mean Sea Level MTI Misleading Terrain Information NASA National Aeronautics and Space Agency NGA National Geospatial-Intelligence Agency NIMA National Imagery and Mapping Agency PDF Probability Density Function
PFFD Probability of Fault-Free Detection PMD Probability of Missed Detection PSD Power Spectral Density RA Radar Altimeter RADAR Radio Detection and Ranging RoMD Region of Missed Detection RTCA Radio Technical Commission for Aeronautics SA Situation Awareness SBAS Space Based Augmentation System SRTM Shuttle Radar Topography Mission SVS Synthetic Vision System TAWS Terrain Awareness and Warning System TIS-B Traffic Information Systems - Broadcast TRN Terrain Referenced Navigation TSO Technical Standard Order USGS United States Geological Survey WAAS Wide Area Augmentation System WGS84 World Geodetic System 1984 WxRadar Weather Radar 15
1
INTRODUCTION
Most aircraft are equipped with instruments to aid the pilot during flight in Instrument Meteorological Conditions (IMC). IMC are characterized by poor or no visibility due to bad weather conditions and darkness. This lack of visibility reduces the number of visual cues available to the pilot and may result in loss of spatial orientation. In order to address the problem of low visibility during a flight and to improve aircraft and airport operations, in general, research on Synthetic Vision Systems (SVS) is being conducted by the National Aeronautics and Space Agency (NASA) in collaboration with various industry and academic institutions, Ohio University being one of them. In this context, this thesis addresses two issues: firstly, development of methods to improve terrain database integrity monitors for use in SVS and secondly, extension of the integrity monitor concepts for application to terrain referenced navigation (TRN).
The basic idea of SVS displays is to provide computer-generated moving images to the pilot that must be a true representation of the outside world. Out of all the information that the computer has to collect in order to generate this ‘outside world’ picture, information about the terrain features is derived on-board from terrain elevation databases. As computers are incorporated in all aspects of flight, the greatest challenge facing avionics researchers is to ensure the accuracy, availability, integrity and continuity of operation of the modern systems. These required 16 navigation performance parameters have to be considered during the design phase and depend mainly on the inherent technologies used. Integrity stems from the component level and depends heavily on the reliabilities of the individual components, which can be studied rigorously using, among others, fault-tree analysis. Terrain database integrity refers to the capability to detect errors in the terrain database that could cause or contribute to the failure of a system function. This thesis deals with improving the integrity of the information derived from terrain databases using sensor information in real-time and estimation techniques such as Kalman filtering and spectral estimation. The methods thus developed will find applications in a Terrain Database Integrity Monitor and for Terrain Referenced Navigation.
1.1 Contributions
The work presented in this thesis elaborates on some of the fundamental work done previously by: Gray, R. A., in his doctoral dissertation titled “Inflight Detection of Errors for Enhanced Aircraft Flight Safety and Vertical Accuracy Improvement using Digital Terrain Elevation Data with an Inertial Navigation System, Global Positioning System and Radar Altimeter”, at Ohio University, Athens, Ohio, in June 1999. [1] Campbell, J., in his masters’ thesis titled “Characteristics of a Real-Time Digital Terrain Database Integrity Monitor for a Synthetic Vision System”, at Ohio University, Athens, Ohio, in November 2001. [2]
The sections of the aforementioned work relevant to this thesis are described in Chapter 3. Some of the material presented in this thesis has previously been published as the conference papers: