An Experimental Study of the Human Interface with One Atmosphere Diving Suit Appendages by Christopher Michael Wilkins B.S., United States Naval Academy (2009) Submitted to the Department of Mechanical Engineering in partial fulfillment of the requirements for the degrees of AASSACHUSETTS INSTITUTE Naval Engineer's Degree OF TECHNOLOGY and JUN 0 2 2016 Master of Science in Ocean Engineering at the LIBRARIES MASSACHUSETTS INSTITUTE OF TECHNOLOGY ARCH-IVES June 2016 @ Massachusetts Institute of Technology 2016. All rights reserved.

Signature redacted A uthor ...... Department of Mechanical Engineering 4ay 3, 2016 Certified by...... Signature redacted Alexandra H. Techet Associate Professor Thesis Supervisor Certified by...... Sig nature redacted CAPT Joel P. Harbour Professor of the Practice of Naval Construction and Engineering A - - Thesis Suuervisor A ccepted by ...... Signature redacted - Rohan Abeyaratne Chairman, Department Committee on Graduate Studies Department of Mechanical Engineering 2 An Experimental Study of the Human Interface with One Atmosphere Diving Suit Appendages by Christopher Michael Wilkins

Submitted to the Department of Mechanical Engineering on May 3, 2016, in partial fulfillment of the requirements for the degrees of Naval Engineer's Degree and Master of Science in Ocean Engineering

Abstract

This experimental study of the human interface with an Atmospheric Diving Suit (ADS) develops a method for quantitatively evaluating how the pilot interacts with the suit's appendages to inform design improvements and to provide a baseline of joint performance for existing technologies. An Atmospheric Diving Suit is a one person anthropomorphic pressure vessel, with manually operated maneuverable appendages, capable of carrying a diver to great depths in the sea while maintaining the internal cabin pressure at one atmo- sphere (14.7psi). Commercial ADS are used regularly around the world in offshore industries, and military ADS are used by a large number of navies for submarine rescue capabilities. This study specifically investigates the performance of the arm rotary joints on the OceanWorks International HARDSUITTM rated for use as deep as 1200 feet of seawater, that are owned and operated by Phoenix International. The experiments were performed at Phoenix International facilities using their own experienced pilots and suits. Experiments were conducted with four different pilots, each performing a series of deliberate, repetitive arm motions while submerged in a shallow training pool. Each pilot was outfitted with a pressure sensor pad placed on the wrist at the major contact region with the appendage, and a series of inertial measurement units (IMUs) placed along the arm and suit. The results of the data analysis show the shape, location, magnitude and move- ment of the contact areas between pilot and appendage as well as peak pressures, dy- namic force loading profiles, impulse and work measurements experienced by the pilots across the specific motions performed. An analysis is performed on the force contri- butions of the hydrodynamic drag acting on the appendage during motion through the application of Slender Body Theory paired with motion data from the IMUs.

3 Thesis Supervisor: Alexandra H. Techet Title: Associate Professor

Thesis Supervisor: CAPT Joel P. Harbour Title: Professor of the Practice of Naval Construction and Engineering

4 Acknowledgments

The completion of this research could not have been accomplished without the sig- nificant support of many contributors. The following organizations and individuals were instrumental in the success of this project: Phoenix International for their enormous generosity in hosting our experiments and contribution of time and resources. Specifically, thanks to Tom Bisset, Terry Breaux, Gary Smith, Erica Lenhart, and Rostislov Haulik who dedicated significant time and effort for this research. The U.S. Navy for support through the STTR and other individuals and com- mands that contributed time and resources for interviews and tours. Thanks to LCDR Jonathon Gibbs, CAPT Keith Lehnhardt, Dr. John Camperman, Mr. Andy Little, and CDR Michael Runkle. The MIT Manned Vehicle Laboratory for guidance on experiment methodology, providing the sensors used in experiments, and training for equipment use and data analysis. Thanks to Dr. Dava Newman, Alexandra Hilbert, and especially Eddie Obropta, who was heavily involved in this research from beginning to end. MIDE Technology for access to their engineers and facilities and guidance for what research contributions would be most useful for advancing ADS technology. And of course a special thanks to my wonderful wife, Lindsay, who took care of Silas, Delia, and our newest addition while I was busy writing this paper.

The author of this paper is a Lieutenant in the United States Navy, an Engineer- ing Duty Officer, and a U.S. Navy Salvage Diver. The author can be contacted at wilkins.c.mOgmail.com

5 6 Contents

1 Introduction 15 1.1 M otivation ...... 15 1.2 Objectives, Aims, Contributions ...... 17 1.3 Thesis Outline ...... 18

2 Background 21 2.1 A Review of ADS Design Evolution Through History 21 2.1.1 Earliest Versions of ADS Technology .... . 21 2.1.2 Transition to Modern ADS Technology . .. . 23 2.2 Present Day Use ...... 25 2.2.1 Commercial Offshore Construction, Inspection, and Maintenance 25 2.2.2 Military Use as Submarine Rescue Capability 27 2.3 Modern Suit Limitations ...... 28 2.3.1 Rang of Motion/Type of Motion ...... 28 2.3.2 M anipulator ...... 28 2.3.3 U.S. Navy Maintenance ...... 29 2.3.4 U.S. Navy Certification ...... 29 2.3.5 U.S. Navy Inspiration for This Research .. . 30

3 Experimental Design and Methodology Used to Identify the User Interface of the ADS 31 3.1 O verview ...... 31 3.2 Setup and Design ...... 32

7 3.3 Sensors ...... 35 3.3.1 Sensor Type ...... 35 3.3.2 Sensor Calibration ...... 36 3.3.3 Sensor Placement and Orientation . 37 3.4 Test Protocol ...... 37 3.5 Lim itations ...... 40

4 Quantification of Human Interaction with ADS Appendage Through Data Analysis 43 4.1 Overview ...... 43 4.2 Peak Pressures .... . 44 4.3 Forces ...... 46 4.4 Inclusion of IMU data . 48 4.5 Discarded Data .... . 52 4.6 Peak Forces And Torques 53 4.7 Calculations of Effort . . 56 4.7.1 Impulse .. .. . 56 4.7.2 Work ...... 59 4.8 Functional Task .... . 63 4.9 Hydrodynamic Drag . . 65

5 Conclusion 73 5.1 Sum m ary ...... 73 5.2 Lessons Learned...... 74 5.3 Future W ork ...... 75

A Additional Data Plots 77

B ADS Test Plan 84

8 List of Figures

2-1 Lethbridge Diving Engine (Image Source: www.therebreathersite.nl, 2016) ...... 22

2-2 Neufeldt and Kuhnke Diving Shell (Image Source: Scott, 1932) ... . 22 2-3 Peress' Tritonia ...... 23 2-4 N ew tsuit ...... 24 2-5 Nuytco Exosuit (Image Source: Nuytco.com, 2016) ...... 25 2-6 Diver in 1200 fsw Hardsuit (Image Source: Phoenix International) . . 26 2-7 Navy Diver in 2000 fsw Hardsuit (Image Source: U.S. Navy) ... .. 27

3-1 Elbow Rotation ...... 33 3-2 Arm Raise ...... 33 3-3 Test Subject Performing Functional Task ...... 34 3-4 Novel Pressure Pad (left) and Hardware Systems (right) (Image Source: Anderson, 2014) ...... 35 3-5 Novel Calibration Tool (Image Source: Novel, 2013) ...... 36 3-6 Test Subject 1 Dressing Out ...... 37 3-7 Graphical Representation of the Sensor Collecting Data: Cell readings and Colorbar units are kPA. The terminal ulna and radius are the bony points of a wrist used to mark the inside and outside of the arm. .. . 38 3-8 Test Subjects Embarking the Suit ...... 39 3-9 Phoenix International ADS Training Pool and ADS Transportation via Crane ...... 39 3-10 Phoenix Dive Supervisor and MIT Investigator in Control Module .. 40

9 4-1 Characteristic Peak Pressure Plot of Both Movement Groups for Each Test Subject ...... 44 4-2 Graphical Representation of the Sensor Collecting Data ...... 45 4-3 Snapshot of Test 3 Elbow Raise ...... 46 4-4 Characteristic Force Plots of Both Movement Groups for Each Test Subject ...... 47

4-5 Representative Force and Arm Angle Plots for Test Subject 1 . ... 49

4-6 Representative Force and Arm Angle Plots for Test Subject 2 ... . 50

4-7 Representative Force and Arm Angle Plots for Test Subject 3 . .. . 51 4-8 Plot of Force vs Time of an Elbow Rotation Motion Group with Peak Forces for Distinct Movements Marked with Red Circles ...... 53 4-9 Plot of Force vs Time with Impulse Area Shaded and Values Labeled in Newtons-seconds for Test Subject 3, Arm Raise Group ...... 57 4-10 Plot of Force vs Time for Each of the Functional Tasks Completed. . 63 4-11 Orientation of Axes Overlaid on the Appendage ...... 66 4-12 Drawing of One Centimeter Wide Strips of the Appendage Used for Measuring Diameters ...... 68 4-13 Calculated Drag Synchronized with Recorded Forces and Arm Angles

for an Arm Raise Motion Group Performed by Test Subject 1 .... 70 4-14 Calculated Drag Synchronized with Recorded Forces and Arm Angles for an Arm Raise Motion Group Performed by Test Subject 2 .. .. 71 4-15 Calculated Drag Synchronized with Recorded Forces and Arm Angles for an Arm Raise Motion Group Performed by Test Subject 3 ... . 72

A-1 Force and Arm Angle Plots for Test Subject 1 Motion Groups 1 ... 77

A-2 Force and Arm Angle Plots for Test Subject 1 Motion Groups 2 ... 77

A-3 Force and Arm Angle Plots for Test Subject 1 Motion Groups 3 ... 78

A-4 Force and Arm Angle Plots for Test Subject 1 Motion Groups 4 . .. 78

A-5 Force and Arm Angle Plots for Test Subject 2 Motion Groups 1 ... 78

A-6 Force and Arm Angle Plots for Test Subject 2 Motion Groups 2 . .. 79

10 A-7 Force and Arm Angle Plots for Test Subject 2 Motion Groups 3 ... 79 A-8 Force and Arm Angle Plots for Test Subject 2 Motion Groups 4 . .. 79 A-9 Force and Arm Angle Plots for Test Subject 3 Motion Groups 1 ... 80 A-10 Force and Arm Angle Plots for Test Subject 3 Motion Groups 2 . .. 80 A-11 Force and Arm Angle Plots for Test Subject 3 Motion Groups 3 . .. 80 A-12 Force and Arm Angle Plots for Test Subject 3 Motion Groups 4 ... 81 A-13 Force Plots for Test Subject 4 ...... 81

11 12 List of Tables

3.1 Test Sequence: Arm raises marked in green, elbow rotations marked in yellow ...... 34 3.2 Test Subject Characteristics ...... 41

4.1 Numbered Motions by Test Subject and Motion Groups. Red cells represent motions for which the force data has been discarded from further analysis...... 52 4.2 Peak Force Mean and Standard Deviations (N) by Motion and Test Subject ...... 54 4.3 Test Subjects Anthropomorphic Measurements for Torque Calculations 54

4.4 Peak Torque Mean and Standard Deviations by Motion (N * m) . . . 55 4.5 Impulse Data (N * s) ...... 58 4.6 Impulse Means and Standard Deviations (N * s) ...... 59 4.7 Work Means and Standard Deviations (J) ...... 61 4.8 Ratio of Impulse to Work ...... 62 4.9 Functional Task Time to Complete and Impulse Calculations by Test Subject ...... 64 4.10 Maximum Estimated Rotary Joint Torques Provided by OceanWorks International ...... 65

A.1 Impulse Data (N * s) ...... 82

A.2 Impulse per Radian Data ( . .) ...... 82 A.3 Work Data (J) ...... 83 A.4 Work per Radian Data ( ) ...... 83

13 14 Chapter 1

Introduction

1.1 Motivation

Human presence in the deep sea is valuable for the advancement of scientific knowl- edge, for expanding the capability of human engineering in the ocean environment, and for protection of national security. As unmanned systems rise to prominence in the age of robotics, the value of maintaining and expanding human access to the harsher environments of the ocean is still critical for providing first person ob- servation, situational awareness, real time adaption to unexpected challenges, and superior dexterity and agility. Humans still surpass robotics at accomplishing all of these things, especially underwater, and can continue to do so with the advancement of technologies and systems that allow humans safe access to the deep. Humans can survive only within a narrow band of environmental conditions. These restrictive conditions include temperature, pressure, light, gas mixtures for breathing and many others. While a human body can enter the ocean without the aid of technology, it is severely limited to shallow depths for short periods in moderate temperatures. There have been many technologies developed over (at least) centuries to extend human range and endurance underwater, which has produced a plethora of diving specialties such as Self Contained Underwater Breathing Apparatus (SCUBA) and saturation mixed gas diving. Nearly all methods utilized for direct interaction between humans and objects in the ocean environment involve subjecting the body

15 to ambient pressure at depth, with all the risks that it entails. Diving as shallow as 33 feet of sea water (fsw) subjects the body to an ambient pressure double that experienced at the surface (one atmosphere). The deepest recorded ambient pressure dives in open water were conducted at depths as great 1750 fsw by commercial mixed gas saturation divers in the Mediterranean in 1988 (Ciesielski and Imbert, 1989). At this depth, divers are subjected to an ambient pressure that is 54 times greater than atmospheric pressure. The U.S. Navy Dive Manual describes the risks of hyperbaric exposure to include a host of physiological disorders that can, and regularly do cause serious and even fatal injuries to divers (NAVSEA, 2011). The single exception to ambient diving methods is a one atmosphere diving suit (ADS) that keeps the human pilot at the same pressure as experienced at the surface.

An ADS is a one person anthropomorphic submarine with articulated appendages. The internal "cabin" pressure is maintained at one atmosphere, protecting the diver, commonly referred to as a pilot, from the perils of hyperbolic exposure. The manu- ally operated articulated appendages surround a pilot's arms, and in most variants their legs as well, like sleeves formed by spherical pressure hull joint sections. The articulated appendages allow the pilot to have a great degree of mobility and direct mechanical interaction with the subsea world without exposure to any increase in pressure.

ADS technology has steadily improved over centuries, but the human interface with the maneuverable appendages remains largely undeveloped. The dexterity, agility, and sensitivity of the appendages are the most valuable parts of the ADS, allowing interaction with the environment that is otherwise inaccessible to humans, yet no study of the ADS has ever been conducted in regards to operability of the user interface. A quantitative understanding of the human interface with the ADS will provide the framework for intentional and demonstrable advancement of human access to the harsh environment of the deep ocean.

16 1.2 Objectives, Aims, Contributions

The objective of this research is to advance the understanding of human interaction with ADS appendages in order to inform the development of improved appendage technology. This understanding is formed through an evaluation of experimental data collected in a series of submerged experiments with pilots in the ADS. The analysis focuses on both the effort exerted by the pilot and the mechanical performance of the joints. The experiments were conducted in collaboration with James Colgary, Naval En- gineer's Degree Candidate, 2016. Colgary's parallel research investigates the range of motions achieved during the same experiment through an analysis of data collected from inertial measurement units (IMUs) placed on the pilot's arm. The two bodies of research overlap in several areas, and portions of data and results are shared between studies, as indicated throughout, to enhance the depth of analysis of both. Jointly generated figures and tables are shared between this paper and his (Colgary, 2016), and an examination of both papers is suggested for a more complete understanding of the human interaction with the ADS. Pilot exertion is quantified in this study through analysis of peak contact pres- sures and forces on the pilots body, and the level of effort experienced by ADS pilots during simple joint movements. This is accomplished with a commercially available pressure sensing system (Novel GmbH, Munich, Germany) utilized to record the con- tact pressures experienced on the pilots' wrists during a series of deliberate appendage movements. The results of the analysis provide:

" The range of pressures that are experienced on the pilot's wrist during ap- pendage manipulation

" The location on the pilot's arm and the magnitude of the contact forces used to effect appendage motion

" The level of effort endured by the pilot during appendage motion (impulse)

17 The mechanical joint performance analysis is accomplished using the same pres- sure sensor data and incorporating movement data collected by the IMUs for Colgary's research during the same motions. The IMU data is used to evaluate the range of motion accomplished by the pilot's effort and to understand how much of the contact force is translated into appendage movement. Finally, the IMU data is used with measured dimensions of the appendage to calculate the hydrodynamic forces acting on the appendage during motion through the application of Slender Body Theory. These theoretically calculated forces are compared to the experimentally observed forces to provide insight on the quantity of effort expended to overcome drag rather than mechanical joint resistance. The results of the analysis provide:

e The portion of the applied contact force that is translated into appendage mo- tion (work)

e The work done on the appendage per degree of movement

e The contribution of hydrodynamic forces to the motion resistance experienced by the pilot

This experimental analysis produces four primary contributions:

1. The first quantitative analysis of human interface with the ADS arm joints

2. A baseline of performance for existing technology

3. A quantitative method for comparing joint performance of new designs against existing technology

4. A theoretical evaluation of the effects of external joint form and its contribution to the forces experienced by the pilot

1.3 Thesis Outline

Chapter 1 introduces the motivations for ADS appendage joint analysis and describes the objectives of this research.

18 Chapter 2 reviews the history of ADS design and how it used today, the limita- tions of current technology and the specific developments that led to this research. Chapter 3 describes the experimental design and methodology used to collect relevant and accurate data for analysis. Chapter 4 describes the analysis of experimental data to quantify the human interaction with an ADS arm joint. A discussion of results and potential errors is included with each step of the analysis. Chapter 5 is a summary and conclusion of the results, and provides recommen- dations for follow-on research to expand upon this work.

19 20 Chapter 2

Background

2.1 A Review of ADS Design Evolution Through History

2.1.1 Earliest Versions of ADS Technology

John Lethbridge, of the United Kingdom, is credited with designing and operating the first version of a suit that maintained atmospheric pressure throughout a dive in 1715 (Thornton, 2000). Lethbridge's "diving engine", as he called it, was made of wood in the shape of an elongated barrel. It was a closed system leaving Lethbridge to breathe his own recirculated breath until the affects of hypercapnia (C02 toxicity) required him to come to the surface to refresh his air supply. His own claims of diving the suit for up to 34 minutes reaching depths of 72 fsw may be exaggerated, but he was able to use the contraption to great success in salvaging treasures and equipment from several wrecks (Thornton, 2000). While Lethbridge's diving engine would maintain approximately atmospheric pressure internally, the diver's arms were extended outside the hull of the contraption and sealed of with a leather cuff. The arms were thus exposed to ambient pressure, but this would only affect the systemic circulation of the exposed limbs rather than the far greater risks that come with hyperbaric pulmonary circulation that involves exchanges of dissolved gas through

21 the respiratory system to the blood stream.

#, t

Figure 2-1: Lethbridge Diving Engine (Image Source: www.therebreathersite.nfl, 2016)

Despite Lethbridge's success, after his death in 1759 there are no records of his diving engine in use or

the successful advancement of ADS technology until

the late 19th century when rudimentary designs of W-0

suits with enclosed articulating arms began to emerge

(Harris, 1994). In 1917 the first fully enclosed ADS with documented successful performance of articulat-

ing arms was designed by Neufeldt and Kuhnke, a Ger-

man firm that built several generations of the suit with

increasing success (Scott, 1932). The Neufeldt and

Kuhnke suits had multiple ball and socket joints for

both arms and legs, separated by ball bearings to with-

in a rubber skirt to iso- stand pressure and enclosed Figure 2-2: Neufeldt and late from the water (Davis, 1955). The suit was tested Kuhnke Diving Shell (Image Source: Scott, 1932) in depths as low as 530 fsw, but mobility of the ap-

pendages was possible only at much shallower depths

(Scott, 1932).

22 2.1.2 Transition to Modern ADS Technology

In 1932 Joseph "Pop" Peress built an ADS with annular cylinder and pist oni joints. His suit, dubbed the Tritonia, was successfully used to visit the sunken Lusitania at over 300 fsw in 1935. piloted 1y Jim Jarret. Despite successful trials with the British Navy, Peress' was unable to find a market for his advanced ADS technology, at least, not until nearly four decades later when he was introduced to the team from Underwater Marine Equipment Limited (UMEL) (Loftas, 1973).

Feb. 20, 1934. . S PERESS 1,947,657

F-lad Av&. *, 1933 2 S..-S et

far

(b onDaig ( J tD g AllA

(a) "Pop" Peress with his Tritonia ADS in 1930 (Iniage Source: Getty Images Imagno)

Figure 2-3: Peress' Tritonia

UMEL recognized that the booming off-shore oil industry needed better options for getting work done at great, depths. With Peress consulting, they modified and improved ADS technology starting with the restoration of the original Tritonia after decades of lay-up in a warehouse. By 1971, UTMEL was able to design and build a highly articulated ADS with an operational depth of over 1000 fsw and a life support system able to last, over 20 hours underwater. This suit, dubbed the JIM suit in

23 honor of the Tritonia pilot named earlier, became the gateway technology for ADS systems to move from obscure one-off suits for the eccentric, to a highly reliable tool for accessing the ocean's depths.

In 1975, Oceaneering acquired BHD Construction, the UMEL subsidiary created to develop the suit, along with all the exclusive rights to JIM suit technology. After successfully marketing the suit to the offshore industry, Oceaneering had 19 JIM suits operational by 1981 (Thornton, 2000).

Graham Hawkes, an engineer working for Oceaneering designed a suit using many of the principle features of the JIM suit but replaced the legs with a cylindrical extension of the torso. Hawkes tried to manufacture and sell his new suit, called the

WASP, through a new company called Offshore Submersibles, but Oceaneering was able to stop their sales and acquire the technology through legal action claiming that

Hawkes had developed the suit while working for UMEL. Oceaneering still builds and operates several generations of the WASP. mostly for offshore oil platform operations (Thornton, 2000).

U.S. Patent OC. 29, 1985 Sheet i of2 4,549.753

(a) French Navy Diver in the Newtsuit (Image (b) Joint Drawings Source: SPL)

Figure 2-4: Newtsuit

24 In 1984 Phil Nuytten, one of the founders of Oceaneering, patented an oil filled rotary joint that became the key component of the HARDSUIT ' (originally called

Newtsuit after its inventor) that was built by Hardsuits International in Vancouver, BC. Hardsuits, now built by OceanWorks, feature five spherically shaped sections for each arm attached to each other and to the suit, in series by five rotary joints. The legs are similarly built with four sphlerical sections and a, boot. The superior mobility of these joints, even at depth, have made Hardsuits the dominant ADS system in use today (Harris, 1994).

Phil Nuytten continues to lead the advancement of ADS technologies with his own company, Nuytco Re- search, that designs and builds a range of deep sub- mnersible craft. His latest innovations in ADS tech- nology have recently been unveiled in 2015 with the Exosuit, which carries over many of the characteris- tics of the Newtsuit but is significantly lighter with a maximum depth of 1000 fsw. Nuytco claims the suit joints perform with far less resistance and the increased dexterity of the innovative design potentially allows for "swim-able" or self propelled configuration. At the time of this righting no verifiable documentation is available to demonstrate the functionality of this new design in Figure 2-5: Nuytco Ex- real operations. osuit (Image Source: Nuytco.com, 2016)

2.2 Present Day Use

2.2.1 Commercial Offshore Construction, Inspection, and Main- tenance

There are a variety of commercial ADS systems in use by companies across the globe. ADS are primarily used for offshore oil projects including construction, maintenance

25 and inspection. While ADS provide unique access for direct human interaction in the deep, they are often used in combination with ROVs, ambient divers or both. ROVs lack perspective, situational awareness and dexterity, but they provide significant mo- bility and strength, and an impressive arrangement of power tools that can multiply the effectiveness of an ADS pilot. Ambient divers face a host of safety risks, lack endurance time on station, and are significantly restricted in movement in the water column, but the dexterity and agility of a wet diver are unmatched by any ADS or ROV. Ambient. divers can be used to set up ADS work sites or provide assistance for particularly delicate maneuvers without exhausting their alloted bottom time, and leave the bulk of the work to be accomplished by the ADS pilots.

Figure 2-6: Diver in 1200 fsw Hardsuit, (Image Source: Phoenix International)

Phoenix International is a prominent operator of ADS systems that participated significantly in this research. Phoenix is a marine services contractor that utilizes a fleet of eight operational Hardsuits rated to 1200 fsw. Phoenix pilots have logged thousands of hours in the ADS working on projects around the world ranging from offshore platform construction to inland hydroelectric dain inspections. They are also contracted for training, operating, and maintaining ADS capabilities for a number of military that utilize ADS systems as described in section 2.2.2, including for the U.S. Navy.

26 2.2.2 Military Use as Submarine Rescue Capability

Several countries own and operate ADS systems in their navies primarily for subma- rine rescue operations. According to the OceanWorks website. the Hardsuit has been used by or sold to Australia, France, Italy, Japan, Russia, Singapore, Turkey, United

Kingdom, and the United States. There are currently no (ocunented incidents in which ADS systems have been used in response to actual casualties., but they are regarded as an important tool in the event that a rapid response is needed for a distressed submarine.

The U.S. Navy owns and operates a small fleet of custom built Hardsuits that were altered by Ocean-

Works to meet strict regulations governing U.S. Navy submersibles found in the System for Certification Pro- cedures and Criteria Manual for Deep Submergence Systems (NAVSEA, 1998). The changes made, includ- ing the use of forged rather than cast, aluminum, also increased the depth rating to 2000 fsw. These suits are referred to as ADS 2000 through the rest of this paper. The four suits are assigned to Undersea Res- cue Command (URC) where the U.S. Navy has con- tracted Phoenix International to help maintain and op- Figure 2-7: Navy Diver in erate them as well as to train the active duty and reserve 2000 fsw Hardsuit (Image sailors assigned as pilots. Source: U.S. Navy)

The ADS 2000 is part of an over-all Submarine Res- cue Diving and Recompression System (SRDRS) housed at URC in San Diego., Cal- ifornia. URC is charged with maintaining a rapid mobilization capability for SR,- DRS to immediately deploy in the event of a, distressed submarine. According to an article in Undersea Warfare magazine describing URC capabilities, "ADS provides a rapid response capability for [distressed submarine] localization and assessment , hatch clearance, and emergency life support stores replenishment. For specific use

27 with SRC [submarine rescue chamber], ADS is one way to attach the downhaul cable to the submarine from the SRC. From there, the. SRC uses the downhaul cable to drive down to the submarine to then mate and complete rescue operations." (de Vera, 2014)

2.3 Modern Suit Limitations

2.3.1 Rang of Motion/Type of Motion

The ease with which the Hardsuit rotary joints turn at depth, and the safe failure mode when the joints break make it the most reliable ADS appendage option available. Yet despite the superior joint mobility, the human body does not naturally move the way rotary joints allow. To reach a 90 degree angle in the elbow, a simple elbow flexion cannot be accomplished in the suit, rather, the pilot has to complete a series of irregular motions to spin the rotary joints into the right-angle position. The cleverly aligned rotary joints do allow a greater range of motion than even the human body can accomplish, but this becomes a liability when the pilots can find themselves twisted into positions from which they cannot recover and could possibly be injured. Even simple tasks become very challenging for those who do not have sufficient experience to recognize the series of motions required to change joint positions.

2.3.2 Manipulator

The standard manipulator used on ADS as the appendage "hands" are pinchers that resemble pliers. They are open and shut via a handle that the pilot operates with their fingers. The pinchers can be locked in place via a thumb dial. The pinchers can pivot a limited amount in any direction from perpendicular to the hull. The motion is translated through the hull of the spherical hand section of the appendage through a mechanical seal. These pinchers cannot accomplish nearly the dexterity or finesse of the human hand and significant experience is needed to operate them proficiently while performing even a simple task.

28 2.3.3 U.S. Navy Maintenance

(Note: The observations of U.S. Navy unique limitations addressed in sections 2.3.3 and 2.3.4 are purely the opinion of the author based on operator interviews and do not reflect official U.S. Navy, URC, or Phoenix International positions.) The U.S. Navy ADS program suffers from a number of self-inflicted capability impairments. Among the most debilitating is the outsourcing of routine maintenance, such as joint rebuilds, to the vendor, OceanWorks. Joint rebuilds, which are routinely performed in-house at minimal cost by Phoenix International on their own 1200 fsw suits, cost the U.S. Navy tens of thousands of dollars and months of downtime while the joints are shipped off-site. The cost and lost time translates to greater periodicity between maintenance and leaves the joints operating with sub-optimal performance. Experienced U.S. Navy pilots complain of joints routinely "locking up" during dives and leaving the appendages immobile. One pilot described a dive to 2000 fsw where all four appendages became locked and he was rendered immobile for several hours while the suit was recovered. In contrast, the Phoenix pilots diving the commercial 1200 fsw suits, which have nearly identical joints but more frequent in-house rebuilds, describe even one joint locking up as a very rare event. A more in-depth examination of U.S. Navy joint lock-up is given in Colgary's thesis (Colgary, 2016).

2.3.4 U.S. Navy Certification

Another ADS operational impairment unique to the U.S. Navy is the extremely con- servative off-gassing certification requirements for humans working in an enclosed submerged space. Far exceeding the off-gassing requirements of the National Aero- nautics and Space Administration (NASA), the U.S. Navy's System Certification Procedures and Criteria Manual for Deep Submergence Systems (P9290) requires testing for all non-metallic materials that enter an enclosed manned space. These tests, which are required for everything from the pilots underwear to electrical wire insulation, cost upwards of $30,000 per sample. At such a high cost, URC can only afford to perform off-gas tests periodically on "batch" samples that include a large

29 number of items being tested at once with no distinction of what item might be the source of contamination if levels are too high. An additional complaint heard from engineers involved in the off-gassing certifi- cation process is that acceptable levels of contamination are derived from industrial standards in which personnel face long-term exposure to the contaminants during 40 hour work weeks over an entire career. This may be reasonable for submarines, but U.S. Navy pilots typically are expected to be in the ADS for training exercises that occur a few times a year, where they are diving every few days, up to 6 hours at a time. Navy sailors are typically only assigned to URC for three to five years before transferring to their next job. To the authors knowledge, there has never been a re- ported incident in an ADS, military or commercial, where off gassing contamination was suspected of causing adverse affects to the pilot.

2.3.5 U.S. Navy Inspiration for This Research

URC is scheduled to phase out the ADS 2000 component of SRDRS and replace it with ROV assets in the next few years due to the unsustainable cost of maintenance and upkeep and the other limitations described in this section. The U.S. Navy diving community recognized a loss of manned deep sea intervention capability, and is look- ing to inspire a new generation of more affordable, agile, deployable, and maintainable ADS technology through a Small Business Technology Transfer solicitation (STTR contract N13A-T029) from the Office of Naval Research (Sullivan, 2013). The STTR was awarded to MIDE Technology in Medford, Massachusetts with the Massachusetts Institute of Technology (MIT) as a University Partner and Phoenix International in Largo, Maryland as a SubContractor. The research described in this paper is con- ducted under the STTR and is intended to aid the development of new ADS joint technology by establishing a quantifiable method for evaluating the performance of new designs against the baseline performance of existing technologies and by inform- ing the design process with insight to the human interface with the appendages.

30 Chapter 3

Experimental Design and Methodology Used to Identify the User Interface of the ADS

3.1 Overview

This chapter presents the process of quantifying the human interface with the ap- pendage joints through a description of the experimental methods used.

The design of this experiment is an adaptation of experiments performed by the Manned Vehicle Laboratory of the Aeronautical and Astronautical Department at MIT with NASA to study the interface between astronauts and space suits. This effort was guided by previous unpublished work on the extra vehicular mobility unit hard upper torso by the Anthropometry and Biomechanics Facility at NASA Johnson Space Center and by multiple MIT thesis research studies on human-space suit interaction

(e.g. Anderson (2014), Hilbert (2015)). These studies focused on understanding the potential for shoulder injury while conducting training in the space suit. The experiment methodology in these studies was altered to focus on the forces engaged with the ADS appendage to effect movement rather than the affect of space suit appendage motions on the human body. While the experiment protocol and data

31 collection is very similar to the space suit experiments, the analysis in this research considers very different aspects of the collected data.

3.2 Setup and Design

As was discussed in section 2.3.5, this research is intended to aid the development of new ADS joint technology by informing the design process and creating a method for evaluating the performance of new designs against the baseline performance of the existing system. The scope of the experiments and analysis conducted were constrained by time and availability of equipment and personnel. Only four pilots were available to serve as test subjects, and the experiment was conducted only once for each subject. The intention of this study is to validate the experimental methodology as an accurate means for evaluating the human interface with the ADS rather than providing statistically significant results. The aim of the study is to support the U.S. Navy goals described in the STTR (Sullivan, 2013). To do this, the commercial 1200 fsw variant Hardsuit was used exclusively for these experiments rather than the U.S. Navy owned ADS 2000 for several reasons. Foremost was the U.S. Navy's cost prohibitive off-gassing require- ments, discussed in section 2.3.4 and outlined in the P9290 (NAVSEA, 1998). The off-gassing tests for the pressure sensors, IMUs and supporting equipment would cost approximately $30,000 with no guarantee of success, a price and a risk that could not be undertaken in the scope of this project. The use of sensors and supporting equipment in the commercial suit was approved based on a comprehensive materials list and NASA's prior approval for use of the same test equipment in space suits dur- ing similar experiments (Hilbert, 2015) (Anderson, 2014). Additionally, the STTR requested a depth capability of 1000 fsw, which is more comparable to the commercial Hardsuit depth rating. Finally, the joint performance on the commercial Hardsuit is superior to the U.S. Navy variants for reasons discussed in section 2.3.3 and therefore provided a better baseline of current joint technology capability for comparison. This is the first quantitative look at the human interface with an ADS appendage,

32 and the priority is to focus on understanding simple movements to ensure fidelity in the analysis. The experiment was designed to demonstrate the level of effort required to move the appendage through simple motions. The motions were restricted to focus on one joint at a time to avoid complicated compound movement that would cause difficulties for analysis.

Two specific movements were chosen to evaluate in this experiment. The first is an elbow rotation focusing on the fourth joint fron tHie shoulder. An elbow flexion motion was desired, however. the comnplex rotary joint motions required to minic an

elbow flexion do not fall into the "simple movement" focus of this experiment. To

perform the elbow rotation the pilots were instructed to move through their full range of motion, forward and back, rotating only the fourth joint. The second motion is an

extended arm raise focusing on the shoulder joint. The pilots were instructed to keep their arm extended and move from a relaxed position down to raising the extended arm above their head and returning to the starting position.

Figure 3-1: Elbow Rotation

Figure 3-2: Arm Raise

Each of these motion types was repeated 20 tiies per test subject in four groups of five up and down repetitions. The order in which the motion groups were performed was altered for each subject by a random sequence with rest breaks between every

33 other motion group as shown in table 3.1. During each rest break the pilots were

questioned for subjective feedback.

After completing the deliberate imotions., the pilots

performed a task with a focus on function rather than form. The functionarl task was designed to demonstrate

motions that would actually be performed by pilots dur-

ing a working dive. The task assigned was to tighten

a nut onto a bolt on a flange. The task was performed using a socket wrench adapted for use with an ADS by

that Figure 3-3: Test Subject attaching a leash to the end of the wrench handle Performing Functional Task was suitable for gripping with the mnanipulators and

could translate the appendage motions into the proper

rotational plane of the socket wrench. A picture of the modified tool and task setup in use by an ADS pilot is shown in figure 3-3.

Task Task Task Task Task Task Task Task Task

arm arm elbow elbow arm elbow arm elbow Subject 1 raisel1 raise 2 rot 1 rot 2 raise 3 rot 3 raise 4 rot44 Subject 21 ~ F 1 Sbet2ebow UL arm U. clow ar U arIlo UU rm ebw F rota 1 raise 2 rot2 I raise 2 raise 3 rot 3 raise 4 rot 4

Sujc3 arm tv elbow 'a arm o arm 'o elbow 'e elbow tv elbow o arm ro Sbet3raiserots I1a1 raroti ise 12 t raiserot 2 2 raise 323 raiserot 24 rot 32 ~E raisrot 443 raiserot 4 4 Q FT

arm Sarm 4 elbow arm arm elbow elbow elbow 4 I ot rs raise FT Subject 1

Table 3.1: Test Sequence: Armn raises marked in green, elbow rotations marked in yellow

34 3.3 Sensors

3.3.1 Sensor Type

Two distinct sensor systenis were used in the exI)eriment to collect data for analysis; pressure sensors and inertial measurement units (IMIU). The primary device used for research covered in this paper was the Pliance® pressure sensor pad purchased from

Novel GmbH (Munich. Germany). The pad used in this study is a modified version of the s2073 that has 128 individual square sensor cells made of capacitive transducers, each 1.4 cm in length and width. arranged in a 16x8 grid. The dimensions of the pad are approximately 22.4 cm in length and 11.2 cm in width. Each sensor cell can record pressures from 5 to 600 kPa and for this experiment runs at a rate of 50 Hz.

The pad connects to a power and on-board storage device through a flexible cable.

The device is powered by 10 1.2V nickel metal hydride batteries producing 2000 mAh and running at 330 mA (Novel, 2013). Data can be transmitted through a fiber optic cable, wireless bluetooth signal, or captured on-board the device on a memory card. Only the on-board storage method was acceptable for this experiment as the device was contained inside an aluminum pressure vessel and submerged underwater. This system or similar versions have been used in the space suit studies described in section 3.1.

Figure 3-4: Novel Pressure Pad (left) and Hardware Systems (right) (Image Source: Anderson, 2014)

The IMUs are APDM Opal Inertial Measurement Units that were used for mea-

35 surenient of both suit and subject kinematics. Their use and resulting data analysis is addressed more thoroughly in parallel research accomplished during the course of the same experiment by Colgary (Colgary, 2016).

3.3.2 Sensor Calibration

Figure 3-5: Novel Calibration Tool (Image Source: Novel 2013)

The Novel pressure pad was calibrated using the purpose built Novel calibration

system consisting of a rubber membrane housed inside a secure unit as shown in

figure 3-5. After the sensor pad is placed inside the calibration tool the rubber

membrane is pneumatically inflated distributing a constant pressure to all the sensors.

The calibrating software prescribes a sequence of pressures to reach while it adjusts

the input data to match.

The calibration can be tailored to the expected pressure range that the sensors

will be capturing. For this experiment, a general calibration was used for the full

range up to 600 kPa because no expected pressure range was established prior to

performing the tests. The calibration equipment is too large to travel with, so no

adjustments were made through tie course of the experiment. In future studies with

the ADS, the sensors could be more precisely calibrated for the 0 to 200 kPa range.

36 3.3.3 Sensor Placement and Orientation

The Novel sensor was placed on the right arm of each test subject with the length of the pad wrapping around the circumference of the wrist. The location of the sensor placement was chosen based on interviews with Hardsuit pilots who described the wrist as the primary contact region for the forces exchanged during arm movements. U.S. Navy pilots report routine bruising on the wrists from the forces experienced while moving the appendages in the ADS 2000. While this is not a common occurrence for the commercial pilots in the 1200 fsw suits, the wrist is still the main point of contact for manipulating the position of the rotary joints.

The Novel data collecting system can only use one sensor pad at a time, and only one pad available had sufficient dimensions to collect data for the full circumference of the arm. The sen- sor does not capture all of the interaction be- tween the suit and the pilots' arms, but the other contact points were described as not con- -!7 tributing significantly to the motion of the ap- Figure 3-6: Test Subject 1 Dressing Out pendage.

3.4 Test Protocol

The tests were conducted in the Phoenix International ADS saltwater training pool in Bayou Vista, LA. The pool is a vertically oriented cylinder with a diameter of 11 feet and a depth of 13 feet. The test pool has a large observation window accessible from the ground. Each pilot that participated in this study operated a Hardsuit rated to 1200 fsw. The suits are adjusted with spacers to fit the individual pilot's size.

The test subjects were fitted with sensors and test equipment in the warehouse ad- jacent to the test pool. Sensors were powered on and calibrated prior to donning the

ADS. While sensors recorded data internally for analysis purposes, they also trans-

37 Elbow Hand * RAnfnr Pad Pr&-zijr& fnicnalu IkPal

Terminal Ulna (Outside Wrist)

underneath arm

around Overlap Wrapped wrist

over arm

Ter minal Radius (Ir side Wrist)

Figure 3-7: Graphical Representation of the Sensor Collecting Data: Cell readings and Colorbar units are kPA. The terminal ulna and radius are the bony points of a wrist used to mark the inside and outside of the arm. mitted data via bluetooth so the investigators could verify systems were functioning properly before transmissions were cut off by the aluminum suits. Once inside the suits, the Phoenix International personnel had full control of oper- ating the ADS systems including life support, communications, and thrusters, as well as transporting the suit and pilot via crane into the test pool. All of these operations were conducted according to their own safety procedures with no interference from the experimental equipment or investigators. After the test subjects were in position in the pool, they were given a fainiliariza- tion period to perform the movements in the suit to prevent an appreciable learning curve skewing the data during the progression of the experiment. The subjects were instructed to perform the tasks in the order presented in ta- ble 3.1 via, the dedicated communications circuits in the suits. All communications with the pilots were relayed through the Phoenix International watch stander in

38 Figure 3-8: Test Subjects Embarking the Suit

Figure 3-9: Phoenix International ADS Training Pool and ADS Transportation via Crane

the control module located adjacent to the test pool. One investigator directed the sequence of events from inside the control module while the other investigators mioni- tored the status of video recording systems and prepared for the breakdown and setup of subject transitions.

Upon the completion of the directed tasks, the Phoenix International personnel extracted the ADS fromn the pool via crane and replaced it in the suit cradle for disembarking the pilot and reconfiguring for the next test.

A detailed test protocol can be found in Appendix B.

39 Figure 3-10: Phoenix Dive Supervisor and MIT Investigator in Control Module

3.5 Limitations

The pressure sensor only collected data from the contact between the suit and the pilots wrist. It is assumed in the following analysis that the contact with the wrist is the only driving force for the observed appendage movements. Some of the pilots describe contact with the appendage in locations not covered by the sensor pad. In some cases it is obvious from the data that the sensor was not recording good contact data during movements. These motions are identified in section 4.5. While the assumiption is known to be inaccurate in some cases, the majority of the collected data during movements provides reasonable results as discussed throughout this chapter.

The study was limited to only four test subjects due to tinie constraints and pilot availability. The subjects evaluated represented a ])road diversity of anthro- sample size pomnorphic characteristics described in table 3.2. Because of the small and large diversity, no meaningful conclusions can be imade about the effect of diver anthropomorphic variability on performance.

The tests are conducted submerged, but not deep enough to provide a substantial hydrostatic pressure. The suits evaluated are rated to 1200 fsw, while the target depth for a new design is given in the STTR as 1000 fsw (Sullivan, 2013). The experiment was conducted at depths shallower than 10 fsw in an above ground pool.

40 Sex Age (yrs) Height Weight (Ibs) Experience (yrs) Subject 1 F 42 5' 7" 142 5 Subject 2 M 47 5'8" 165 17 Subject 3 M 40 6'2" 220 8 Subject 4 M 50 5'5"1 180 10

Table 3.2: Test Subject Characteristics

The expense of conducting experimentrs at greater depths is well beyond the budget for this research. It is expected that the joints would operate with stiffer resistance at greater depths, however. the Phoenix international pilots interviewe(d for this research did not recognize an appreciable difference in difficulty for maneuvering appendages

at depth. Several of the Navy pilots did cite increased joint resistance with depth

in the ADS 2000. which is believed to have worse performance than the commercial

Hardsuit as described in section 2.3.3.

41 42 Chapter 4

Quantification of Human Interaction with ADS Appendage Through Data Analysis

4.1 Overview

This chapter follows a progression of analytical complexity to describe the process of understanding the interactions recorded by the sensors in the experiments detailed in Chapter 3. Starting with the basic pressure sensor output, the data is visualized, summed, integrated, and statistically examined. Results and data from the IMU analysis described in Colgary's thesis is then synchronized with the pressure sensor data and incorporated in a more complex analysis of energy translated into appendage motion through the application of force from the pilots wrist. Finally, a theoretical estimation of hydrodynamic drag is calculated using Slender Body Theory to pair the dimensions of the appendage with the velocities and accelerations recorded by the IMUs to determine its contribution to the forces recorded by the pressure sensor.

43 4.2 Peak Pressures

The simplest way to graphically analyze the pressure data is to look at the largest pressure value recorded by any sensor cell at each time interval (the sensor samples at 50 Hz). The plot of peak pressure over time displays the trend of contact intensity during appendage movements. A peak pressure plot of a representative motion group is shown in figure 4-1 for each pilot and motion type. These plots only show the highest value of the 128 cells recorded by the sensor at each time interval, and does not give any insight as to the location, shape or magnitude of the contact interaction.

Arm Raise Groups Elbow Rotation Groups Test Subject 1 Test Subject 1 -*200

100

{- 0- 0 1 2 3 4 5 6 7 E 9 10 11 12 13 14 15 16 17 18 19 0 1 2 3 4 5 b 7 9 9 10 11 12 13 14 15 Test Subject 2 Test Subject 2 210 c200 100 100 r 0 - 0 1 2 3 4 5 6 7 9 9 10 11 12 13 14 15 16 17 18 192021 22 0 1 2 3 4 5 7 1C 1 1 12 13 14 15 16 17 Test Subject 3 Test Subject 3 200 O 200

100 7 CLi100

0 1 2 3 4 5 6 7 8 9 1011 1213141516171819202122232425 26 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2021 22 Test Subject 4 Test Subject 4

200 200

100 100 20 0 - 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Time (s) Time (s)

Figure 4-1: Characteristic Peak Pressure Plot of Both Movement Groups for Each Test Subject

In order to clarify what the contact interaction looks like, the pressure data is dis- played on a color map representing the sensor cell grid. The orientation of the sensor pad on the pilot is shown earlier in figure 3-7. The images shown in figure 4-2 present a series of contact shapes and magnitudes observed for each subject performing the different motion types. Each image shows an instantaneous look at the sensor data

44 that, is representative of the contact made in each repetition.

Test Arm Raise Arm Raise Elbow Rotation Elbow Rotation Color Subject Orientation Up Down Up Down (kPa)

Orientation Key Body Mark

Terminal Ulna (outside wrist) 2 Sensor Pad Overlap

Discarded daia

*Datadiscarded 3 due to pinch in the sensor pad causing ;-2 irrelevant noise, _ or redundancy from overlap

Figure 4-2: Graphical Representation of the Sensor Collecting Data

It is clear from figure 4-2 that each test subject performs the motions with unique contact locations. The participants were only directed with the desired appendage joint performance, and were left to their own preference for how to manipulate the suit. This variation introduces another layer of complexity in analyzing the per- formance of the suit but does allow for greater insight, into how experienced pilots interface with the appendages. The lack of uniformity in how the appendage is ma- nipulated suggests that there is no clear optimal technique, which could be a, valuable place to focus for design improvements or even technique training. An improved in- terface design might contribute to better dexterity and less muscle fatigue for the

45 pilots even without a change to the joint, mechanism.

The imagery also makes clear that the sensor

Sensor Pad Pressure Dispaly (kPa) pad does not capture the entire contact region,

NaN and in some cases, significant portions of the pad NaN

325 are rendered useless because of wrinkles or pinchies NaN that caused the cells to recordl abnormally high

NaN pressures. These factors are addressed later in sec- 325 N N data to discard for 725 25 tion 4.5 when discussing which

N~N nunmerical analysis. NaN

NaN The shape and intensity of the contact region

NaN can also give insight into how pilots develop bruis-

NaN ing while operating the suit, a problem particu-

NaN larly common in the ADS 2000 variant. As shown NaN NaN NaN NaN NaN NaN NaN

7, tj in figure 4-3, the pilots often experience high pres-

Figure 4-3: Snapshot of Test 3 sures in a localized region which could lead to Elbow Raise bruising. An improved interface design might dis-

tribute the pressure over a greater area and reduce the potential for injury from operating appendages.

4.3 Forces

The peak pressure plots only show values from one sensor cell per time interval, and the graphical pad display shows the data from each cell but only for one time interval. The contributions of the entire pad can be displayed throughout the whole movement period by converting the pressure data into force data. To do this, the recorded pressure reading from each cell is multiplied by its area on the grid, which is uniformly 1.96cm2 , and the sum of each sensor cell contribution is calculated for every time step. The force values are critical for evaluating the performance of the joint as they give insight into what level of effort each pilot, exerts on the appendage to affect motion.

46 I -Cl

Figure 4-4 shows the dynamic force loads recorded by the sensor pad for each pilot in a representative motion group for both arm raises and elbow rotations. These plots show another dimension to the variance between pilots in how they perform the motions from the smoothness of the curves to the distinction, or lack there of, between up and down motions. Peak forces experienced during a motion also change significantly between pilots with ranges of 20 to 120 N (5 to 30 lbs). Significant variance is even evident between individual motions, which can be observed in the nearly 200% increase in peak forces from start to finish for test subject 1 arm raises. However, there are some similarities as well, such as the shape and intensity of the force loading between the arm raises for both test subjects 2 and 3.

Arm Raise Groups Elbow Rotation Groups Test Subject 1 Test Subject 1

z- 10 . - 0-z 1 2 I 01 21 41 0 - ~- 0 1 2 3 4 5 6 7 V 9 10 11 12 13 14 15 16 17 18 19 0 1 2 3 4 5 7 P 0 10 11 12 13 14 15 Test Subject 2 Test Subject 2

50

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 192021 22 10 0 1 2 3 4 56 7 t 10 11 12 13 14 15 16 17 Test Subject 3 Test Subject 3 0- 103 100

IL LLL 0 0 1 2 3 4 5 6 7 8 9 1011 1213 11 1516 1718192021222324 2526 0 1 2 3 4 5 6 7 8 9 1011 12 13 14 1516171819202122 Test Subject 4 50 z- Test Subject 4

100

50 LL ( tr

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Time (s) Time (s)

Figure 4-4: Characteristic Force Plots of Both Movement Groups for Each Test Sub- ject

When compared to the peak pressure plots of the same movements in figure 4- 1, it is clear that the peak pressure magnitudes do not correlate perfectly with the magnitude of the forces. This is because tile calculated force is a summation of all the contributing pressures recorded on the sensor, rather than just the highest value.

47 The force plots are generally smoother and show a clearer picture of what is occurring in the interaction between suit and pilot during the motion. The variation of force magnitudes, even across a single motion group, suggests that the sensor pad is not observing the entire contact region during motion. A comparison of the force plots with the sensor pad snapshots of the same motion groups shown in figure 4-2 does show a correlation between inconsistent force magnitudes and pressure groupings immediately adjacent to the edge of the senor. There is almost certainly a portion of the contact region that migrates off the sensor pad during these movements.

4.4 Inclusion of IMU data

From the force measurements alone, it is impossible to precisely identify the periods of motion of the appendage or determine in which direction the motion is occurring. To do this, it is necessary to include the data from the inertial measurement units (IMUs) shared from Colgary's research. After the IMU data was filtered and analyzed as described in Colgary's research (Colgary, 2016), the resulting relative angles between IMUs were time synchronized with the Novel pressure sensor data to delineate the individual motions. The IMU data for test subject 4 was found to be unusable because of noise corruption from an unknown source. For this reason, data from test subject 4 will not be addressed in any analysis performed with the inclusion of IMU data. For the arm raises, the relative angles between IMUs placed on the chest and the forearm of each test subject is used to determine the movement of the appendage. For the elbow rotations, the relative angles between IMUs placed on the bicep and the forearm of each test subject are used to determine the movement of the forearm portion of the appendage. In both cases it is assumed that the base IMU (chest for arm raises, bicep for elbow rotations) remains stationary during the movements. This is a reasonable assumptions based on observation of video recordings of the motions. The time stamps of the angle maximums and minimums are used as start and stop times for up and down motions. An up and down motion combined form an individual repetition of the motion type. As described previously, there are five individual

48 111111111M!

repetitions in a motion group. The following figures demonstrate the synchroilized movement groups with the angle minimums marked with magenta lines and angle maxinums marked with green lines. The interval between the magenta and green lines is considered the "up" period of the motion, while the interval between the green and magenta lines is considered the "down" period.

Arm Raise Group Elbow Rotation Group Force Force

130 -11 I

P50 5')

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1t 19 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Time (si T ime (s)

Arm Angle Arm Angle

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 0 1 2 3 4 5 6 7 B 9 10 11 12 13 14 15 Time (s) Time (s)

Figure 4-5: Representative Force and Arm Angle Plots for Test Subject 1

The plots for test subject 1 arm raises show that the sensor consistently captured the interaction between pilot and suit appendage for the upward motion, however, recorded forces during the downward motions are largely nonexistent. These down- ward motions are discarded from the data for the purposes of numeric evaluation because the sensor is clearly not sufficiently capturing the contact region. The elbow rotation with time stamps marked provides an unexpected outcome showing that the pilot rolled from the upward motion to the downward motion without, breaking contact with the suit appendage. It is notable too that the peak forces register during the arrested motion at the peak of the relative angle. This is possibly a result of the rapid transition from upward to downward motion reversing the direction of momentum. Figure 4-5 only shows one group of each motion type performed by test subject 1. The other motion groups can be found in Appendix A and they produce similar trends that show consistency across the experiment. Test subject 2 shows a very consistent arm raise force profile in the movement

49 Arm Raise Group Elbow Rotation Group Force Force

103 1

0 1 2 3 5 6 7 8 9 10 11 12 13 14 15 16 17 18 192021 22 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Time (s) Time (s)

RepresentativeA4-6: Force and Ar Angle Plots for Test Subject 2 group demonstrated in figure 4-6. This consistency is repeated throughout the other arm raise motion groups for test subject 2 that are shown in Appendix A. The sinmilaritv in force magnitude and profile for both the upward and downward motions, as well as the location of the contact patterns on the sensor padl display as shown in figure 4-2 suggests that this data is likely very accurate for Plirpo5ss of evaluating the human-suit interface and joint performance. The clean breaks in the force plot that align nearly perfectly with the motion time stamps also increase confidence in the accuracy of evaluating the effort expended during each motion. The consistency of the force profile does not carry over to the elbow rotations for test subject 2. It is clear from figure 4-6 that virtually nmo force is recorded during the downward motions of the elbow rotations, a result that is repeated with few exceptions in the motion groups that can be found in Appendix A. Apart from the missed data for the downward motions, the force proflles show a consistency in the shape of up~ward motions, but vary greatly in magnitude.

An interesting theme ini the elbow rotation upward motion force proflle is the two peaks with a drop to zero in between. It appears in the graphical pad display that the contact force is truly decreasing to zero at this point rather than migrating off the padl. This suggests that dlurilig this brief p~eriodl of stable rotational sp)eed that the app~end~age momentum is carrying the motion on until it is arrested by the regained contact that is represented by the second peak.

50 g Arm Raise Group Elbow Rotation Group Force Force

1 1 2 3 4 5 t 7 8 9 101112131415161718192021222324252i 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 15 19 20 21 22 Time (si Time (s)

Arm Angle Arm Angle

C 0 1 2 3 4 5 6 7 8 9 1011 121314 151617181920212223242526 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2021 22 Time (s3 Time (s)

Figure 4-7: Representative Force and Arm Angle Plots for Test Subject 3

The arm raise force plot displayed in figure 4-7 for test subject 3 has a similar consistency as was found for test subject 2. There is a notable difference, however, in the shape of some of the up motions from the follow-on down motions that are also found in the other motion groups shown in Appendix A. As with the data from test subject 2, the consistency of shape and magnitude suggests a high accuracy in the data for numerical evaluation. It is notable that the peak forces observed here are greater in magnitude than those observed for test subject 2.

The force profiles of the elbow rotations shown in figure 4-7 demonstrate a consis- tent pattern that appears to capture the contact interaction between pilot and suit for both the upward and downward motions. A closer inspection of the graphical sensor pad display shown in figure 4-2 does show that the downward motion contact area is directly adjacent to the cells that are disregarded due to overpowering noise.

It is likely that a portion of the contact interaction is lost in the noisy cells. However, the shape of the force profile as well as the shape of the contact region on the pad display suggests that the majority of the contact region is captured by the sensor and likely only a peripheral portion of the contact interaction is lost.

This pattern of good data capture is not repeated consistently in the other motion groups for test subject 3, which can be found in Appendix A. Only in a few other instances are the downward motion forces clearly represented. It is likely that in these cases the contact region migrated off the working portion of the sensor pad and were

51 m

not recorded.

4.5 Discarded Data

After a careful review of the force plots compared to the relative angle plots described in section 4.4 and an examination of the contact shape, location and migration (luring the motions, certain portions of the force data were determined to have insufficiently captured the contact interaction to be included for numerical analysis. These motions unfortunately make up a significant portion of the experimental data but carry too much known error to be useful for providing insight into the interaction between suit and pilot. Table 4.1 shows the motions that had inadequate data for further analysis marked in red.

Test S bject I Test S bject 2 Test S bject 3 Test Subject 4

Aw-. nc... .-3,OArdRaswr -A

1 1 _ 1 1 1 3 1 5 3 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 4 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 3 5 5 5 5 5 5 5 5 5 5 5 5 1 1 1 1 3 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 5 _ _ 5 5 5 5 5 5 5 5 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 - 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 L 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 1 1 1 1 1 3 1 1 1 1 1 ,. 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 5 5 5 5 5 55 . 5 5 5 5 5 5 5 5 5 Table 4.1: Numbered Motions by Test Subject and Motion Groups. Red cells represent, motions for which the force data has been discarded from further analysis.

The disregarded motions represent one third of the experiments conducted and are disproportionately weighted towards downward motions for both arm raises and elbow rotations. The disregarded downward motions for test subject 4 are misleading because the IMU data for test subject 4 was unusable, and the distinction between motions is unclear from the pressure sensor data alone. The force profiles are shown in Appendix A and typically only show five distinct impulses for arm raises or an

52 I incomprehensible profile for the elbow rotations. It was decided to consider the dis- tinct peaks as up motions to be included only for peak force analysis as discussed in section 4.6. For all other analysis which requires accurate synchronization with the appendage movement, data from test subject 4 is not considered.

4.6 Peak Forces And Torques

The peak forces experienced during movements are the moments with the highest interaction between the pilot and the appendage. To understand the extent to which this interaction occurred during the experiments, the peak forces for each movement were collected and analyzed for characteristic distinctions between test subjects and motion types. Figure 4-8 shows a representative motion group with the peak forces of each motion marked.

Force 100i 1 TI* T I

o I --Ij

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Time (s)

Figure 4-8: Plot of Force vs Time of an Elbow Rotation Motion Group with Peak Forces for Distinct Movements Marked with Red Circles

Table 4.2 shows the averages of peak forces broken up by movement type and test subject. The table also gives the standard deviation of these peak forces by group. The standard deviation is then divided by the mean, which normalizes the variation by displaying it as a percentage of the mean, making it easy to compare with other standard deviations. Only the motions marked in blue in table 4.1 were included in the calculations for this data. Clearly, the peak forces recorded throughout the experiment are not repeating with consistency in magnitude. Even the most consistent motion, the arm raise down (which is skewed by the bulk of data that was disregarded), has a standard deviation of 25% of the mean.

53 Test Subject: 1 2 3 4 Combined Mean (N) 39.5 30.8 48.8 93.7 53.4 a $ o(N) 9.7 6.5 14.2 11.2 26.6 o/Mean(%) 25% 21% 29% 12% 50% E Mean (N) 34.5 31.4 34.0 N/A 32.9 S(N) 14.4 8.4 6.0 N/A 8.1 o/Mean(%) 42% 27% 18% N/A 25% Mean (N) 125.6 56.9 75.9 119.4 93.5 I 0 (N) 33.6 20.0 16.4 54.3 42.4 I a/Mean(%) 27% 35% 22% 45% 45% C Mean (N) 92.6 N/A 38.4 N/A 71.3 11 N 15.9 N/A 7.8 N/A 29.6 o/Mean(%) 17% N/A 20% N/A 42%

Table 4.2: Peak Force Mean and Standard Deviations (N) by Motion and Test Subject

An examination of the combined average peak contact forces shows that the elbow rotation up motion forces are nearly three times greater than those experienced during the arm raise down motion. To understand the disparity in peak forces across motion types and even test subjects, it is important to remember that not only are the movements being accomplished in different joints, but also that torque measurements give a more complete picture of the mechanics of rotary joints. To convert the force measurements into torque measurements, the force must be multiplied by the distance of the applied force from the center of rotation. In this case, the distance is from the center of the rotary joint (assumed to be pilot's shoulder for arm raise, pilot's elbow for elbow rotation), and the location of the contact area (assumed to be at the center of the pressure pad on the wrist). The distance used to make the conversion are shown in table 4.3, which lists the applicable anthropomorphic measurements of the pilots.

Distance Measured Subject 1 Subject 2 Subject 3 Subject 4 Shoulder to Pad 0.66 (m) 0.65 (m) 0.61 (m) 0.57 (m) Elbow to Pad 0.25 (m) 0.26 (m) 0.23 (m) 0.25 (m)

Table 4.3: Test Subjects Anthropomorphic Measurements for Torque Calculations

54 The averages of the peak torques, grouped by motion type and test subject, along with the associated standard deviations and percentages are listed in table 4.4. The difference between the combined average peak torques for arm raise down and elbow rotation up motions is negligible. This dramatic change in magnitude comparisons from table 4.2 to table 4.4 demonstrates the importance of examining the same data set through different units of measurement. The force measurements provide insight to the performance of the person and potential causes for injury, bruising and fatigue, while the torque measurements are useful for understanding the performance of the joint.

Test Subject: 1 2 3 4 Combined Mean (N*m) 26.0 20.0 29.8 53.4 32.4 5 a (N*m) 6.4 4.2 8.7 6.4 14.3

a - /Mean(%) 25%21% 29% 12% 44% E c Mean (N*m) 22.7 20.4 20.8 N/A 20.7 a (N*m) 9.5 5.5 3.6 N/A 5.2 a o/Mean(%) 42% 27% 18% N/A 25% Mean (N*m) 31.4 14.8 17.5 29.9 23.0 .3 a (N*m) 8.4 5.2 3.8 13.6 10.7 a/Mean(%) 27% 35% 22% 45% 46% c Mean (N*m) 23.2 N/A 8.8 N/A 17.5 a (N*m) 4.0 N/A 1.8 N/A 7.7 a/Mean(%) 17% N/A 20% N/A 44%

Table 4.4: Peak Torque Mean and Standard Deviations by Motion (N * m)

While the torques maintain the same standard deviation percentages as the force measurements, the difference in magnitudes across motion groups are significantly reduced. This is an expected result that lends credibility to the accuracy of the data. It also gives clarity to which motions are suspect for inaccuracies, such as the elbow rotation down movements for test subject 3 which are far below the other aver- ages, indicating the sensor may not have captured the complete contact interaction. Recording above average torques, such as the arm raise up movements for test subject 4, may be the result of inefficient maneuvering technique rather than inaccuracies in the data collection method.

55 4.7 Calculations of Effort

To quantitatively measure the effort required to complete a motion, it is necessary to consider the forces engaged with the appendage during the entire motion. Two distinct units of measure are used to evaluate this effort; impulse and work. The impulse is discussed first in section 4.7.1 because it is calculated using only the data collected by the Novel pressure sensor. Section 4.7.2 will describe the calculation and analysis of work values which incorporate the data shared from Colgary's research.

4.7.1 Impulse

Impulse is the integral of the force with respect to time during the period of contact as shown in equation 4.1. Impulse is typically used to evaluate change in momentum over a very small time period, but here it is used to give weight to the duration that contact forces are engaged between pilot and appendage during each motion. It is important to note that the impulse value does not depend on motion of the object, only the contact forces exchanged. For this reason it is useful to calculate the impulse to understand the effort exerted by the pilot, even when that effort is not cleanly translated into motion of the object. It is the sustained effort exerted by the pilot that leads to fatigue rather than the motion of the object itself. Thus, the impulse values provide a more complete picture of the pilot's effort than the work values discussed in section 4.7.2.

t2 Imp = F(t) - dt (4.1) tl where:

Imp =Impulse exerted on the object

t =Time

F(t) =Force acting on the object

The impulse described in equation 4.1 is for a continuous force that changes as a

56 function of time. The sensor data provides a discrete variable force recorded at a spe- cific sampling frequency (50 hZ). The impulse exchanged in the contact interaction for each individual up and down motion is calculated by summing the recorded forces be- tween the motion time stamps and dividing by the sampling frequency as described in equation 4.2. This is essentially taking the area under the force curves of each motion as represented in the drawing in figure 4-9. Because of the difficulty of establishing accurate motion time stamps without the IMU data, the impulse calculations for test subject 4 were not included in this analysis.

t2 Imp Z F - dt (4.2) i=tl where:

F =Force acting on the object at time step i

dt =Sample rate (1//50Hz)

Force

2 4 53 57 6 54 48 48 49 59

0 1 2 3 4 5 7 8 9 10 11 12 13 1 15 16 17 18 19 20 21 22 2.' 24 25 26 imve (s)

Figure 4-9: Plot of Force vs Time with Impulse Area Shaded and Values Labeled in NewTtons-seconds for Test Subject 3, Arm Raise Group

As was discussed in section 4.5, a close examination of the force data was per- formed with the aid of the IMU data and graphical representations of the pressure interaction to determine if the sensor was sufficiently capturing the contact interac- tion. For the motions that were deemed to have insufficiently captured the contact interaction, the impulse data was discarded and is represented in table 4.5 with blank spaces. The impulse, being dimensionalized by time, has the effect of normalizing the force readings to the period it is engaged. For example, the impulse for higher forces M!Y!!P!FP

Test SubJect 1 Test S bject 2 Test Subject 3 Arm Raise Elbow Rotation Arm Raise Elbow Rotation Arm Raise Elbow Rotation

____ up down up down U p down u p down uIp down u p down 115.8 30.5 60.8 60.0 42.8 52.9 56.3 17.7 43.6 30.8 20.6 36.7 43.2 57.1 62.6 28.5 18.2 32.1 34.9 28.4 19.7 17.3 53.9 48.2 33.0 37.6 53.5 33.9 14.9 31.6 30.8 56.7 47.6 19.9 32.9 49.5 37.5 33.0 31.5 48.8 58.6 93.6 14.3 25.1 22.6 60.0 39.1 29.5 36.8 20.7 50.4 58.0 55.8 2_ 36.8 12.2 32.1 37.0 32.1 23.2 61.0 43.6 54.7 C ~ 33,5 17.2 61.1 23.8 25.1 13.8 47.5 52.1 78.9 37.1 32.1 21.0 58.1 26.1 26.6 64.4 46.8 41.8 18.1 2 26.6 45.7 33.4 51.9 28.4 42.5 63.4 33.9 65.9 23.7 11.9 69.6 54.3 50.4 29.0 40.3 35.4 63.5 20.0 0 20.5 49.3 40.5 57.4 29.4 27.5 31.5 73.5 29.0 0 C 22.7 44.2 47.1 58.7 31.8 24.6 37.5 29,0 61.5 23.8 34.5 45.6 49.0 48.6 38.7 38.7 42.1 27.7 58.5 15.7 43.2 56.4 51.0 53.1 36.5 26.6 46.3 46.5 65.3 22.4 20.3 107.8 30.2 34.2 34.7 40.5 41.3 58.3 23,9 49.2 43.8 36.3 32.3 57.4 57.1 47.3

C 19.1 39.2 33.3 38.5 24.8 58.2 69.8 64.7 62.2 13.5 19.5 40.6 81.8 43.0 32.8 31.5 78.9 82.1 60.5 13.3 2 23.1 1 47.0 55.5 58.4 35.3 71.8 65.1 42.5 19.2

Table 4.5: Impulse Data (N * s) recorded during a rapid motion are comparable to the smaller forces recorded over a longer period in a slower motion. This is demnonstrated in the data given in tables 4.5 and 4.6 where there is consistency in movement effort across both test subjects and movement types, even with the wide variation in movement strategies, contact lo- cations, and force profiles. The right most column in table 4.6 shows the combined averages of all three test subjects in each motion type. Each mean value listed in this column is within one standard deviation of all the others. This means that de- spite all the variation in how the movements are accomplished, each notion type was performed with a similar range of recorded impulse effort. These results provide a valuable basis for understanding how much effort is required to manipulate the articulated appendages.

For purposes of practical comprehension, a 40 N * s impulse experienced by a pilot is equivalent to holding a 4.5 pound weight for 2 seconds An analysis of impulse divided by range of mnotion was conducted using the relative angle minimums and maximums of each motion. The results are not discussed here because they are inconclusive with standard deviations sometimes even greater than

58 the mean. A table of the means and standard deviations of this analysis can be found in Appendix A.

Test Subject: 1 2 3 Combined Mean (N*s) 26.5 40.1 52.9 40.1 * N a (N*s) 8.4 13.3 12.6 15.9 o/Mean(%) 32% 33% 24% 40% E Mean (N*s) 17.3 33.2 49.2 40.0 a (N*s) 4.2 8.4 13.6 15.0 o/Mean(%) 24% 25% 28% 37% Mean (N*s) 53.7 32.6 56.1 48.8 a (N*s) 21.6 13.2 16.8 20.5 o/Mean(%) 40% 40% 30% 42% = Mean (N*s) 40.9 N/A 19.4 33.0 a (N*s) 13.4 N/A 4.8 15.2 o/Mean(%) 33% N/A 25% 46%

Table 4.6: Impulse Means and Standard Deviations (N * s)

4.7.2 Work

Calculating the work performed on the appendage during motions is a means for determining the effort that actually translated into movement of the joint. Work is defined by Merriam-Webster as "the transference of energy that is produced by the motion of the point of application of a force and is measured by multiplying the force and the displacement of its point of application in the line of action" (merriam- webster.com, 2016). The applicable mathematical definition of work is shown in equation 4.3.

s2 t2 W = F -ds = T.w - dt (4.3) s ti

59 where:

W =Work exerted on the object

F =Force applied

s =Distance of the object along the line of displacement

t =Time

T =Torque acting on the object

w =Angular velocity of the object

The latter portion of equation 4.3 was utilized to determine the work performed on the joint to achieve movement. The force measurements described in section 4.3 were converted into torque measurements in the same process described in section 4.6 utilizing the distances from center of rotation to the application of force that are presented in table 4.3.

The angular velocity of the appendage was calculated using the data from the IMUs placed on the forearm of each pilot, shared from Colgary's research, using equation 4.4. It is assumed for these calculations that the movement of the appendage is equivalent to the movement of the pilot's arm, and that the magnitude of the velocity (speed) is sufficient for determining the work.

W = +W2 +W2 (4.4) where:

w =Angular velocity of the object

WX, w,, w =x, y, z components of angular velocity from IMU data

The IMU data was smoothed by taking the average of each sample with its ad- jacent samples because the angular velocities were not clean and the sample rate for the IMU data was nearly three times that of the Novel pressure sensor (128Hz to 50Hz respectively). The torque data for each time step was then multiplied by the

60 closest-in-time smoothed angular velocity magnitude before being multiplied again by the force sample rate of 1/50(s). Each of these values was summed for the duration of movement as described in equation 4.5.

t2 W = ( Fi - d -w,(round(i . fIMU/fNovel))dt (4.5) i=tl where:

d =Distance to applied force from table 4.3

WS =Angular velocity smoothed by averaging

fNovel =Frequency of Novel pressure sensor sample rate (50Hz) fIMu =Frequency of IMU sample rate (128Hz)

The results of the work calculations are shown in table 4.7. It is important to recognize that, even though work and torque are dimensionally equivalent and torque is used in the equation for calculating work, they are completely different concepts and are not interchangeable. The work data presented is labeled in joules (J) while torque data is presented in N * m to avoid confusion.

Test Sub'ect: 1 2 3 Combined Mean (J) 24.0 21.5 26.7 24.1 G. 2 (J) 8.5 6.7 6.8 7.7 g o/Mean(%) 36% 31% 26% 32% E Mean (J) 9.7 22.9 25.1 23.0 o J) 2.0 6.9 6.3 7.5 c/Mean(%) 21% 30% 25% 32% Mean (J) 12.5 5.9 13.4 11.0 3 c (J) 5.2 2.4 3.6 5.1 o/Mean(%) 42% 41% 27% 47% C Mean (J) 6.5 N/A 7.4 6.8 o (J) 2.9 N/A 2.0 2.6 a/Mean(%) 44% N/A 28% 39%

Table 4.7: Work Means and Standard Deviations (J)

The work values calculated are a representation of how the applied force is trans- fered into kinetic energy of the appendage. While the impulse values provide insight

61 to the level of effort required from the pilot, the work values provide insight into how much of that effort is translated into motion. The averages in the Combined column of table 4.7 show that the elbow rotations require less than half the work calculated for the arm raises. This is an interesting difference from the impulse data in table 4.6 where there was a close grouping of magnitudes across movement types. The ratio of impulse to work from the Combined columns of tables 4.6 and 4.7 are presented in table 4.8. This ratio can be used as a relative measure of movement efficiency, where a lower value indicates higher efficiency because a greater part of the applied force is being transfered into motion.

Arm Raise Elbow Rotation Up Down Up Down 1.67 1.74 4.44 4.85

Table 4.8: Ratio of Impulse to Work

Similar impulse values were found for both arm raise and elbow rotation move- ments, however, the elbow rotations in general completed a far smaller range of mo- tion. In addition, the applied force for the elbow rotations is much closer to the center of rotation, meaning that the distance traveled by the point of contact (which is key to the work calculation) is much smaller than the arm raise, even for an equal angu- lar rotation. These factors, along with any other contributers, such as the awkward motions required by the human arm to effect a elbow joint rotation, are the likely cause of the lower movement efficiencies for the elbow rotations.

An analysis of work divided by range of motion was conducted using the calcu- lated angle rotations of each motion (E" w -dt). The results are not discussed here because they are inconclusive with standard deviations sometimes even greater than the mean. A table of the means and standard deviations of this analysis can be found in Appendix A.

62 4.8 Functional Task

The final exercise for each test subject was to perform a functional task of tightening a nut onto a flange using an adapted socket wrench, as described in section 3.2. The exercise was designed to reflect an actual job an ADS pilot might perform on a working dive. The force profile plots for the functional tasks are displayed in figure 4-10. Test subject 1 was unable to complete the task after the socket wrench became dislodged. The precision novemeints required to reset the wrench proved too difficult for the pilot and the exercise was abandoned. This failure underscores the limitations of the ADS joints, even for simple tasks. For this reason there are no functional task results to discuss for test subject 1.

Functional Task, Test Subject 2

60

0!

0 10 20 30 40 50 Time (s)

Functional Task, Test Subject 3 153'

100v

0 10 20 30 40 50 60 Time (s) Functional Task, Test Subject 4

300 lII -

LL 200 I i' %V\ A

0 L7 0 10 20 30 40 50 60 Time (s)

Figure 4-10: Plot of Force vs Time for Each of the Functional Tasks Completed.

Each pilot was directed to perform the task as if they were on a working dive with no thought to what motions are being used. This was specifically designed to examine how the pilots use the appendages outside of the unnatural repetitive motions that are assigned for the rest of the experiments. The result is that the pilots used complex

63 motions involving multiple joints rotating simultaneously to complete the task. The IMU data was not fully analyzed for these exercises due to the complexity of motions and was subsequently not shared for inclusion in this research. For the purposes of this research, the only analysis performed on the functional task is an examination of the force profiles shown in figure 4-10 and the calculation of impulse for each pilot shown in table 4.9.

Subject: 2 3 4 Time to complete: 50 (s) 70 (s) aborted at 55 (s) Impulse: 835 (N*s) 2530 (N*s) 5990 (N*s)

Table 4.9: Functional Task Time to Complete and Impulse Calculations by Test Subject

As expected for a free form task, a wide variation is observed in force profile shapes, peak forces experienced, and time taken to complete the task. Test Subject 2 demonstrates a consistent repetitive motion approach with little change from start to finish on the force profile. Understanding what is happening in the force profiles of test subjects 3 and 4 requires a review of the video and communications recordings of these exercises. Test subject 3 maintains a consistent pattern on the plot until around the 42 second mark the sensor goes quiet for about 10 seconds. The video shows that at the same time the subject loses grip of the wrench and takes the 10 seconds to recover his grip and continue the task. Test subject 4 struggles to find a rhythm for the motion and around the 25 second mark loses grip of the platform with his nonworking hand that is being used to stabilize the suit. The rest of the time shown on the plot is the pilot attempting to continue the task while the suit is drifting free in an almost neutrally buoyant state. The exercise is called off around the 55 second mark and is left incomplete. The impulses recorded during the completion of the tasks are shown in table 4.9. These numbers are informative but are not directly comparable to each other because the number of turns on the nut was neither controlled nor strictly monitored. The video shows that the angles through which the socket wrench moves vary greatly

64 between subjects which can explain peak force and profile shape differences.

4.9 Hydrodynamic Drag

To understand joint performance, two characteristic forms of appendage motion re- sistance must be considered. First is the mechanical friction resistance of the joint as it rotates. The second is the fluid forces acting on the appendage during motion. OceanWorks International, the manufacturer of the Hardsuit, provided estimates of the rotary joint torques for this research, which are shown in table 4.10. The hydro- dynamic drag must be calculated using recorded data from the experiments.

Joint: Shoulder Elbow Maximum Estimated Torque: 18 (N*m) 10 (N*m)

Table 4.10: Maximum Estimated Rotary Joint Torques Provided by OceanWorks International

The drag force acting on a cylinder moving through a fluid is (Triantafyllou, 2014):

dU 1 D = ma d++ pCaU|UIL (4.6) dt 2 where:

D =Drag force acting on the cylinder

ma =Added mass of the cylinder U =Velocity of the cylinder

p =Density of the fluid

Cd =Coefficient of drag a =diameter of the cylinder

L =Length of the cylinder

For this study, the drag force is calculated for the purpose of estimating the fluid dynamic contribution to the recorded forces during the experiment motions.

65 --9%!XM%!!Md5CI--M

The drag forces are only calculated for the arm raise motions because they have the entire appendage moving through greater velocities than the elbow rotations and an analysis of the greater contribution is sufficient for demonstrating the extent of the hydrodynamic forces. The inertia drag constitutes the first terms in the drag equation, maI . To cal- culate the added mass coefficient, the appendage is analyzed using Slender Body Theory. Slender Body Theory allows the 3 dimensional added mass coefficients of an object with significant length in one direction relative to the other two directions (slender body) to be estimated using a summation of 2 dimensional added mass co- efficients along the object's length (Techet, 2005). In this case, the entire appendage is considered a, slender body with the 1 axis oriented along the length of the arm and the rotational motion in the 5 direction is induced by the force applied at the wrist acting in the 3 direction as shown in figure 4-11.

6 3

5

W

1

Figure 4-11: Orientation of Axes Overlaid on the Appendage

The added mass force can be described as (Techet, 2005):

Fj= -Uimji - FijU4 2'miT (4.7)

66 for i = 1,2,3, 4,5,6 and j,k,l = 1,2,3

For rotation in the 5 direction only, and with symmetry assumptions, the only remaining non-zero term is:

F3 = - 5m 35 (4.8) where:

F3 =Inertia force in the 3 direction

U5 =Acceleration in the 5 direction

M3 5 =Added mass in the 3 direction due to a unit acceleration in the 5 direction

To calculate M35, a Slender Body Theory equation is used from the book Principles of Naval Architecture (Lewis, 1989):

L

M = xia33idx (4.9) where:

a33 , =ith 2 dimensional added mass in the 3 direction due to a unit acceleration in the 3 direction L =Length of the body

xi =Distance from the center of rotation to the ith strip

dx =The width of each strip along L

To determine the 2 dimensional contributions, the appendage is divided into one centimeter strips along the 1 axis (dx = .01 m) that are assumed to be circular in the 2-3 plane. Diameter measurements for each strip were taken from a scale drawing of the appendage shown in figure 4-12. For simplification, only one drawing was used for all three subjects. The impact of appendage spacers and minor variations of joint positions are assumed to be minimal. The 2 dimensional added mass of each strip was calculated as:

a33 i = p4 di (4.10)

67 ___-- -~ ------

where:

kg p =Density of the water (1000 3

di =Diameter of the ith strip

Figure 4-12: Drawing of One Centimeter Wide Strips of the Appendage Used for Measuring Diameters

Using equation 4.9, the added mass term of the appendage acting against the pressure pad during arm raise accelerations was calculated to be 11.27 kg * m. The contributions of the pincher at the end of the appendage were ignored because of it's small sized and complex shape, as well as the entire first joint from the shoulder because it essentially remains in place during rotation. The acceleration in the 5 direction was calculated by taking the derivative of the angular velocities described in equation 4.4. The errors introduced by the as- sumptions used in calculating the angular velocities, as described in section 4.7.2, are compounded in the acceleration values. In order to produce reasonable acceleration curves, both the angular velocities and the accelerations were smoothed using moving average filters in MATLAB. The smoothed acceleration data is then multiplied by the forces. M 3 5 term for each time iteration to determine the contribution of inertial The later terms in equation 4.6, !pCdaU UJL, express the form drag acting in the opposite direction of motion. To calculate the form drag contribution, the same

68 measured diameters described in equation 4.9 are used. For each time iteration of the smoothed angular velocity data, the form drag is calculated using the summation shown in figure 4.11. L F = - PCddi (WjXi)2dx (4.11) i=r1 where:

Fj =The form drag acting on the body at time j L =Length of the body

p =1000 k (Density of the water)

Cd =1 (Coefficient of drag of a cylinder (Triantafyllou, 2014))

di =Diameter of the ith strip

wj =Smoothed angular velocity of the appendage at time j

xi =Distance from the center of rotation to the ithstrip

dx =0.01m (The width of each strip along L)

The combination of both inertia forces and form drag gives the total drag force acting on the appendage during motion as shown in equation 4.6. This drag force, along with the mechanical resistance of the joint is what the pilots must overcome to manipulate the appendages. The drag forces on the appendage are calculated using the velocity data from the IMUs and the appendage dimensional measurements described above. These theoretically calculated forces are compared to the recorded experimental forces in figures 4-13, 4-14, and 4-15. It is important to remember that form drag is always acting opposite to the direction of motion, but inertial forces act opposite to the direction of acceleration. This means that in the event of a rapid deceleration, the total drag force may be contributing to the motion rather than acting against it, as is seen in the following figures.

The theoretically calculated forces shown in figure 4-13 for test subject 1 provide insight to what the sensor pad failed to detect in the downward motions. The theoret- ical forces are routinely higher in magnitude than the recorded forces. This suggests

69 Recorded Force (Pressure Pad)

60 Z40

U 1 2 3 4 5 6 7 8 9 1i 11 12 13 14 15 16 17 18 19 Time (si

Calculated Drag (IMU) I -1 T T 1 I

I - T -

S1 2 3 4 5 6 7 8 0 11 12 13 14 15 16 17 18 19 Time iS) B\\' Recorded Arm Angle(IMU)

50 0

0 1 2 3 4 5 6 7 6 9 10 11 12 13 14 15 16 17 18 19 Time (s)

Figure 4-13: Calculated Drag Synchronized with Recorded Forces and Arm Angles for an Arm Raise Motion Group Performed by Test Subject 1 that the sensor pad did not capture the full contact region even for the upward mo- tions, as the actual forces exchanged are expected to be greater because they must also overcome the mechanical resistance of the joint. An interesting phenomenon that occurs in this comparison is the negative drag calculated for the deceleration of downward motions appears to correlate with a registered force on the sensor pad. This suggests that the pilot actively arrested the momentum of the appendage before transitioning to upward movements.

The theoretical forces shown in figure 4-14 for test subject 2 have similar shape and magnitude as the recorded forces. This agreement reinforces the credibility of the experimental measurements and demonstrates that a significant portion of the effort required to move the appendage is spent against hydrodynamic drag. The high level of noise in the drag calculations even after significant filtering of the velocity and acceleration data may be a result of using the IMUs attached to the pilot's arm, which does not directly translate to the movement of the suit. Small adjustments of

70 I Recorded Force (Pressure Pad)

2

-L - i - I I I JI -- 3 4 f, 7 8 10 11 12 13 14 19 16 17 1P 19 2) 21 22 23 24 Time (s)

Calculated Drag (IMU)

43 -

20

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 16 17 18 IC 20 21 22 23 24 1 lmcI Recorded Arm Angle(IMU) 5D

/ TI

0 1V

2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time (s)

Figure 4-14: Calculated Drag Synchronized with Recorded Forces and Arm Angles for an Arm Raise Motion Group Performed by Test Subject 2 the arm inside the suit would register as unreasonably fast motions for the appendage. The appendage IMU data was not made available for this research.

Much of the characteristic shape of the recorded force plot represented in figure 4- 15 for test subject 3 is reflected in the theoretically calculated force plot, which also agrees well in force magnitude.

The comparison of calculated drag plots synchronized with the recorded force plots does not show enough difference in the magnitudes of the two for a clear conclusion on the contribution of the mechanical joint resistance beyond that it appears to be insignificant. As discussed earlier in section 3.5, the increased pressure at depths where the ADS is typically operated is expected to cause greater frictional resistance to joint rotation. It is likely that if the experiments were conducted at a greater depth, the contribution of mechanical resistance would be much more pronounced.

Conducting human subject testing in an ADS at depth is cost and risk prohibitive, but the hydrodynamic interactions with the suit do not change with ambient pressure

71 Recorded Force (Pressure Pad) 83~~ T I~ 6J P 23

-20 U-. j ------.

.1 -2 I L. I

1 3 4 5 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Time (s)

Calculated Drag (IMU) 80 -I I 60 40

P7 20~ L 0 -- -20 -

1 2 3 5 F 7 110 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Time (s) Recorded Arm Angle(IMU) T 7

0 -

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Time (s)

Figure 4-15: Calculated Drag Synchronized with Recorded Forces and Arm Angles for an Arm Raise Motion Group Performed by Test Subject 3

increases. Conducting unmanned joint mobility tests in a pressurized test facility to determine if an appreciable increase in joint resistance occurs would provide a useful supplement to this study.

72 Chapter 5

Conclusion

5.1 Summary

Atmospheric diving suits are a unique tool for safe and interactive human access to deep waters. They provide a niche capability that has proven to be worth the costs and logistics required to own and operate. ADS are well used in both commercial and military applications. Advancement of the technology is a necessary step to improve performance, drop costs, and allow the ADS to be used more broadly by more people. The design evolution of ADS technologies has, until now, focused mainly on en- gineering improved joint designs. Understanding of the human interface with those designs has been largely speculative and experiential. The results from this research are the first step in developing a better understanding of the inner working of an ADS in operation. The contributions put forward in this thesis are:

1. The first quantitative analysis of human interface with the ADS arm joints

2. A baseline of performance for existing technology with results for:

" The range of pressures that are experienced on the pilot's wrist during appendage manipulation

" The location on the pilot's arm and the magnitude of the contact forces used to effect appendage motion

73 " The level of effort endured by the pilot during appendage motion (impulse)

" The portion of the applied contact force that is translated into appendage motion (work)

" The work done on the appendage per degree of movement

" The contribution of hydrodynamic forces to the motion resistance experi- enced by the pilot

3. A quantitative method for comparing joint performance of new designs against existing technology

4. A theoretical evaluation of the effects of external joint form and its contribution to the forces experienced by the pilot

Quantifying the human interface with an ADS is a difficult task because of the small space available for equipment, the isolated nature of a pressure vessel submerged in water, and the complexity of the interaction between pilot and suit. While the motions and calculations performed are straight forward and simple, it is hoped that they will form the building blocks for more detailed and complex understanding of ADS technology that will lead to improved designs.

5.2 Lessons Learned

There were many unexpected difficulties, errors, and areas for improvement discovered during the experiment that are described throughout this paper. These documented lessons learned provide the opportunity for future investigators to improve their pro- cesses and results for any follow-on research. Some of the key lessons learned are given here with references to where they are described in this paper.

* The pressure can be calibrated for better accuracy in the 5 to 200 kPa range.

(sections 3.3.2, 4.2)

74 " The location of the contact interaction is not captured completely by the sensor used in this experiment wrapped around the pilots wrist. Use of larger sensors, multiple sensors, or better training and awareness on the part of the pilot could improve results. (sections 4.2, 4.3, 4.5)

" The functional task would be more useful if it were better controlled to make it repeatable and comparable. (section 4.8)

" Synchronization of pressure and movement data is essential to understand the nature of the contact interaction. Video cues proved unreliable for time syn- chronization because the cameras used could not record the entire course of the experiment due to battery and overheating issues. The IMU data was necessary for identifying the moments when the applied force was translated into motion.

(sections 4.4, 4.3, 4.7.2)

5.3 Future Work

The methods used in this research only provide the first quantitative look inside the ADS while it is being operated. To better develop the understanding of the human interface with the ADS, future research could be conducted in any of the following areas:

* Repeat the experiments described in this paper to increase the pool of partici- pants and provide enough data for accurate statistical analysis to be performed focusing on the effects of pilot experience, anthropometry, and movement strate- gies.

" Introduce more complex motions for experimental analysis with multiple joints rotating simultaneously.

" Examine the hydrodynamic drag contributions by more accurate paring of IMU and force data during motions performed through a deliberate range of speeds.

75 * Determine the relationship between increased ambient pressure acting on the appendage (increased depth) and the mechanical resistance of the joint during rotation. As described in section 4.9, even unmanned experiments would be a valuable supplement to this study.

76 Appendix A

9 Additional Data Plots I

Arm Raise Group Elbow Rotation Group Force Force

51U 50 0

0 I 0 - 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1V 19 1 2 3 4 5 6 7 B 9 10 11 12 13 14 15 Tir-e ( , Time (Si Arm Angle Arm Angle

100- 100

0' 01 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Time (s) Time (s)

Figure A-1: Force and Arm Angle Plots for Test Subject 1 Motion Groups 1

Arm Rais. Group Elbow Rotation Group Force Force

100 5 100t

U 50' P 50

I 01 0 0 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 16 19 0 1 2 3 4 5 6 74 0 10 11 12 13 14 15

1Ib'I (3 Tire is,

Arm Angle Arm Angle 100 710 50

S1 2 3 4 5 6 7 6 9 10 11 12 13 14 15 16 17 18 19 0 1 2 3 4 5 6 7 0 3 10 11 12 13 14 15 Time (s) Time (s)

Figure A-2: Force and Arm Angle Plots for Test Subject 1 Motion Groups 2

77 p m11110 - -

Arm Raise Group Elbow Rotation Group Force Force

10 - 15 2 * 0 j I' (fi ~ ''-4 AJ f 7 4 0 1C 11 12 1- 17 18 i 1 6 7 1, 12 13 14 I me (S Time (18 Arm Angle Arm Angle

tIC . 44 ~5o. 0

2 3 4 5 6 7 8 0 10 11 12 13 14 15 16 17 18 1 C 1 2 3 4 5 6 7 6 0 C 11 12 13 14 Time (s) Time (s)

Figure A-3: Force and Arm Angle Plots for Test Subject 1 Motion Groups 3

Arm Raise Group Elbow Rotation Group Force Force I.

41 .4, 2 50, ii'

0 0 5 6 7 8 9 10 11 12 13 11 15 16 17 18 14, 5 6 7 8 9 13 14 Time (s Tin (s

Arm Angle Arm Angle

50, 50 C i2 01234 5 6 7 8 91 1 1213141 16 17 18 19 1 2 3 4 5 6 7 8 10 11 12 13 14 Tin (si Time (si

Figure A-4: Force and Arm Angle Plots for Test Subject 1 Motion Groups 4

Arm Raise Group Elbow Rotation Group Force Force

50' U.

LL1 2 -- 2 3 4 5 6 7 5 9 10 11 12 13 14 15 1617 18 19 20 21 22 2 1.3 4i5i 7 8 9 10 11 12 13 14 15 16 17 Time (s) Time (s)

Arm Angle Arm Angle 10 I c , , - 54 - 50 - 12 3-2 35

oi

0 1 2 3 4 0 6 7 8 5 10 1112 13 14 15161718 1920 21 22 D 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Timu (,

Figure A-5: Force and Arm Angle Plots for Test Subject 2 Motion Groups 1 I

78 I Arm Raise Group Elbow Rotation Group Force Force

-100 - 109

50 1 50 0 A' - 0 1 2 3 4 5 6 7 0 9 1 11212114 150117 18192021222324 0 1 2 3 4 5 6 7 P 9 10 11 12 13 14 15 111-5 Arm Angle Arm Angle

1004 100

50'

-n 0'

0 1 2 3 4 5 6 7 8 9 1011121314151617 1812021222324 1 2 3 4 5 6 7 8 9 1C 11 12 13 14 15 Tinw (St Tm (Si

Figure A-6: Force and Arm Angle Plots for Test Subject 2 Motion Groups 2

Arm Raise Group Elbow Rotation Group Force Force

z

250, LL 0 - -X- 50 - 0 1 2 3 4 5 6 7 8 9 10111213 14151617 18192021222324 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 11911< IS

Arm Angle Arm Angle

100 50'

01 0'-

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 212223 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Time (s, Time (s)

Figure A-7: Force and Arm Angle Plots for Test Subject 2 Motion Groups 3

Arm Raise Group Elbow Rotation Group Force Force

100 -1001

501 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1516 17 18 19 20 21 22 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 imA (l3) Tot (tl Arm Angle Arm Angle

51, 10k 010

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 192021 22 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Time (si Tire (Sj

Figure A-8: Force and Arm Angle Plots for Test Subject 2 Motion Groups 4

79

-~ ~ I Arm Raise Group Elbow Rotation Group Force Force

100 106

-2 50 , 50

7 3 13 111121314 15161718192021222324252( 7 11 1' 12 13 14 1 ' 1 '' 2Z1 2223 Tin (si lime is) Arm Angle Arm Angle

103 ~ Si /~ \ 31 01

2 3 4 5 6 7 e 9 1011 121314 151617 11020212223242526 [ 1 2 3 4 5 6 7 e 9 10 11 12 13 14 15 16 17 16 19 2i 21 22 23 Time (si Time (s)

Figure A-9: Force and Arn Angle Plots for Test Subject 3 Motion Groups 1 I

Arm Raise Group Elbow Rotation Group Force Force 10M1 100 i L 1 2 0 ; 0 23 4 5 6 7 8 9 101112131415 1617181920212223242526 3 6 7 8 9 1D 11 12 13 14 15 16 17 18 19 2021 22 Time (si Timew (s) Arm Angle Arm Angle

50 13' 'V "V 7t~\ I 0212

0 1 2 3 4 5 6 7 8 9 10111 21314 1516 1716192021222324 2526 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Time (s) Time (si

Figure A-10: Force and Arm Angle Plots for Test Subject 3 Motion Groups 2

Arm Raise Group Elbow Rotation Group Force Force

100 100

1 2 3 4 5 6 7 8 9 101112131415161716192021222324:25 C 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Time (s) Time (s) Arm Angle Arm Angle

1ao 1W0

50

0 1 2 3 4 5 6 7 8 9 1011121314 1516171819202122232425 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2021 22 Tim(u (13 Time (s)

Figure A-11: Force and Arm Angle Plots for Test Subject 3 Motion Groups 3

80

1I O-ORN mmmmmr_-

Arm Raise Group Elbow Rotation Group Force Force

100I 100' V50 h l~ i ~ .i >

18 19 20 21 2 2 4 6 7 8 1 1112131415161711l2-21222 2.252627 1 5 6 7 8 9 1) 11 12 13 14 16 17 Trs (si Tit I- Arm Angle Arm Angle 5102 50 \ 50~ 0 0 0 < 2 0 1 2 3 4 5 6 7 8 9 101112131415161718192221222324252627 0 1 234 5 6 7 8 1 111 12 1314 10 1617 18 19 2021 Time (s) Time (si

Figure A-12: Force and Arm Angle Plots for Test Subject 3 Motion Groups 4

Arm Raise Groups Elbow Rotation Groups Motion Group I Motion Group 1

100k 10

P, 50 50 0k U-

01 7 8 0 1 2 3 4 5 6 7 8 9 17 11 12 13 14 15 16 17 18 19 20 2 3 4 5 6 9 10 11 12 13 14 15 Motion Group 2 Motion Group 2

10D 200 50

0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 1 2 3 4 5 6 7 1011 1213141516171819202122232425 Motion Group 3 0 Motion Group 3 U- 100 -- 200

50 F2 100 ii - 2 3 0 0 1 2 34 5 h 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 2122 Motion Group 4 Motion Group 4

/oo-/ 200

50 0 2 - 0- 161 01 23 4 56768 9 10 11 12 13 14 15 16 17 18 19 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 5 16 17 18 19 Time (s) Time (s)

Figure A-13: Force Plots for Test Subject 4 I 81 Test Subject 1 Test Subject 2 Test Subject 3 Arm Raise Elbow Rotation Arm Raise Elbow Rotation Arm Raise Elbow Rotation up___I down up dowin up down up dlovn U p down u p down 115.8 30.5 60.8 60.0 42.8 52.9 56.3 17.7 43.6 30.8 20.6 36.7 43.2 57.1 62.6 28.5 18.2 32.1 34.9 28.4 19.7 17.3 53.9 48.2 33.0 37.6 53.5 33.9 14.9 31.6 30.8 56.7 47.6 19.9 _32.9 49.5_ 37.5 33.0 31.5 48.8 58.6 93.6 14.3 25.1 22.6 60.0 39.1 29.5 36.8 20.7 50.4 58.0 55.8 36.8 12.2 32.1 37.0 32.1 23.2 61.0 43.6 54.7 33.5 17.2 61.1 23.8 25,1 13.8 47.5 52.1 78.9

0 37.1 32.1 21.0 58.1 26.1 26.6 64.4 46.8 41.8 18.1 _ 26.6 45.7 _33.4 51.9 28.4 42.5 63.4 33.9 65.9 23.7 0- 11.9 69.6 54.3 50.4 29.0 40.3 35.4 63.5 20.0 20.5 49.3 40.5 57.4 29.4 27.5 31.5 73.5 29.0 Cm 22.7 44.2 47.1 58.7 31.8 24.6 37.5 29.0 61.5 23.8 34.5 45.6 49.0 48.6 38.7 38.7 42.1 27.7 58.5 15.7 _ 43.2 _ 56.4 51.0 53.1 36.5 26.6 _____ 46.3 46.5 65.3 _22.4 20.3 107.8 30.2 34.2 34.7 40.5 41.3 58.3 23.9 49.2 43.8 36.3 32.3 57.4 57.1 47.3 C: 19.1 39.2 33.3 38.5 24.8 58.2 69.8 64.7 62.2 13.5 19.5 40.6 81.8 43.0 32.8 31.5 78.9 82.1 60.5 13.3 0 2 23.1 47.0 55.5 58.4 35.3 71.8 65.1 42.5 19.2

Table A.1: Impulse Data (N * s) I

Test Subject: 1 2 3 Combined Mean (N*s/rad) 9.8 18.8 25.0 17.9 a (N*s/rad) 3.7 7.5 7.9 9.1 aa/Mean(%) 38% 40% 32% 51% E r Mean (N*s/rad) 0.9 11.8 23.5 12.1 o(N*s/rad) 2.2 5.7 9.4 11.3 o/Mean(%) 247% 48% 40% 93% Mean (N*s/rad) 50.6 19.0 65.3 45.0 - a (N*s/rad) 25.7 16.6 50.9 39.4 o/Mean(%) 51% 88% 78% 88% 3 c Mean (N*s/rad) 39.3 N/A 8.6 16.0 0 . o a (N*s/rad) 24.6 N/A 8.6 22.6 a/Mean(%) 63% N/A 99% 141%

Table A.2: Impulse per Radian Data (*)

82 I Test Subject 1 Test S bject 2 Test S bject 3 Arm Raise Elbow Rotation Arm Raise Elbow Rotation Arm Raise Elbow Rotation Li___p down up down up down up down up down up down 29.3 14.2 42,6 5.7 25.0 27.1 13.1 15.2 8.4 3.2 12.1 24.9 7.4 23.6 29.2 5.9 18.7 10.8 5.0 16.1 15.2 2.8 26.4 22.8 6.8 32.9 11.0 5.8 8.4 26.0 4.9 27.4 21.2 5.9 28.2 15.3 4.5 19.6 4.4 22.7 29.1 16.0 4.8 27.6 12.2 10.5 7.2 15.8 22.6 5.8 25.8 33.1 11.1 39.4 7.3 8.0 5.2 18.2 5.3 25.4 19.7 11.7 C N 32.0 9.5 10.2 3.6 15.0 2.9 19.5 24.0 19.0 37.3 7.0 2.4 25.0 12.7 4.9 30.9 22.2 9.9 6.1 23.6 9.1 5.5 21.4 13.0 7.4 30.2 18.7 14.4 8.8 10.6 19.9 7.7 26.3 19.7 22.5 18.8 16.4 6.9 17.6 10.1 6.8 26.3 19.8 14.3 17.7 14.8 11.2 19.0 17.0 7.3 28.7 24.1 5.0 19.4 16,9 14.3 8.0 27.9 10.7 7.2 25.5 26.6 9.9 20.6 16.1 15.4 6.7 38.1 15.1 6.8 28.5 30.8 5.1 22.0 24.5 15.7 10.8 16.3 17.8 5.2 23.1 24.4 26.5 24.3 17.3 19.6 9.0 7.7 21.8 22.1 37.0 32.1 13.7 4 15.9 8.7 4.8 23.1 17,3 12.3 35.1 34.3 14.5 5.1 15.0 9.2 15.4 23.1 23.0 4.3 43.3 36.9 16.2 5.8 21.2 12.0 11.3 38.2 25.2 36.1 33.9 15.2 7.4

Table A.3: Work Data (J)

Test Subject: 1 2 3 Combined Mean (J/rad) 10.9 11.7 15.2 12.6 i a (J/rad) 3.2 3.5 3.8 4.0 a/Mean(%) 29% 30% 25% 32% E Mean (J/rad) 4.1 13.0 22.3 17.0 0 a (J/rad) 1.7 4.5 20.7 16.0 o/Mean(%) 43% 35% 93% 94% Mean (J/rad) 14.4 4.2 9.4 9.8 Sa (J/rad) 5.1 1.5 2.6 5.4 a/Mean(%) 35% 35% 28% 55% SMean (J/rad) 4.1 N/A 15.9 8.4 0 i . o a (J/rad) 2.3 N/A 20.4 13.7 a/Mean(%) 55% N/A 128% 163%

Table A.4: Work per Radian Data (d)

83 I 84 Appendix B

ADS Test Plan

Human-Suit Interface Pressure Evaluation

MIT Ocean Engineering and Phoenix International Diving Division Location: Bayou Vista, LA

Preparedby Jim Colgary, Chris Wilkins, and Eddie Obropta, MIT

1 Introduction The objective of this research is to develop an understanding of how the person interacts with the One Atmosphere Diving Suit (ADS), and use that information to inform future design. Our approach is to quantify and evaluate human-ADS interaction with pressure sensing mats, focusing on the elbow joint. Additionally, inertial measurement units (IMUs) will be placed both on the ADS pilot and to the ADS arm to assess biomechanics. This study builds from previous collaboration between the MIT, NASA, and the David Clark Company and the evaluation of space suit design. Establishing a precedent and proof of concept for this methodology concerning ADS testing will open the doorway for future collaboration, additional suit testing, and technology development.

2 Test Summary Subjects will be asked to perform the test protocol in the OceanWorks ADS (commercial variant: 1200ft). Subjects will be selected based on availability of Phoenix personnel who appropriately volunteer and meet the standard medical requirements for suit operations. These individuals have a great deal of experience piloting the ADS so will not have to develop new, potentially confounding movement strategies. The subjects will be wearing the pressure sensing and IMU systems while performing the tests, and pressure profiles and angle histories will be recorded. The test protocol will consist of 20 repetitions of 2 motions inside the ADS and one function task. The selected movements focus around the elbow, where the sensors are placed. Specifically, there are 2 isolated "typical" ADS movements (elbow rotation/arm raise) and one functional task (ratcheting. Prior to the test, movements will be discussed with the pilot. For each movement, the 20 repetitions will be further subdivided into 4 groups of 5 repetitions each. This is done to evaluate subject fatigue or potential change of biomechanical strategies over the course of the test period. After each group of movements, qualitative information on subject comfort and hot spots will be collected. The information will also be collected after training. Each of these test conditions will be counterbalanced and randomized. 85 3 MIT Hardware The human-suit interface is currently an unknown in ADS characterization. Pressure measurements would allow greater insight into how these interactions occur and help characterize suit performance. Additionally, data collected will allow us to inform future design by creating an "effort of movement" baseline. There is currently no method by which to characterize the human interaction with ADS.

3.1 Novel Pressure Sensing System One Novel pressure sensor will be placed on the pilot's RIGHT arm at 2 inch above the wrist. Note, the Novel pressure-sensing mat has been used previously in a study by the MIT Aeronautical Engineering Department in the evaluation of Space Suits.

- Two commercially available Novel pressure mats, S2073 with 128 sensors each - Each sensor is 1.4cm in each dimension and pressure range between 20-600kPa e Mat placed around RIGHT arm at the wrist covered with a protective sleeve. + On-board data collection with electronics mounted at the base of the back e Sensor runs at 330mA current - Battery is 10 1.2V nickel metal hydride - The system is certified to the European safety standard 93/42/EEC (Annex IX). - Due to the construction of the battery it is unlikely that water or sweat will come into contact with the electronics board of the battery pack. It is also unlikely that a small amount of moisture would create an electrical short-circuit. - The pedar NiMH 2000mAh battery pack is internally secured with an overheating and an overcurrent protection (Polyswitch). e Worst case scenario for puncture: On the transmitter side of the sensor mats a voltage of 7 V (effective) = equal to 20 Vpp is applied (pp = peak to peak). The maximum current for a shortcut is 100 mA(pp) if one directly touches the transmitter. For technical applications the resistance of the human body is typically considered to be 1 - 2.4 kOhm. In that case the maximum current would be 8 - 20 mA(pp).

Figure 2: Novel Pressure sensing system. A) Sensor mat. B) Data acquisition and battery

3.2 APDM IMU Sensing System Additional information about the human-suit interface may be gathered using IMU data collected inside the ADS. There is a joint angle difference between the person's movement and that of the ADS. This is due to the resistance of the ADS to movement, as well as (in some instances) anatomically inaccurate rotation due to bearing movement. Calculating the joint angles measured

86 UAMME6 __Q

on the pilot and of the ADS would help elucidate these differences. As stated previously, this experiment uses similar methods to those used in Space Suit evaluation and were found to be effective in identifying human-suit interaction. This data will be used not only to determine biomechanical differences, but also to help find points of maximum and minimum movement to analyze the pressure profiles. Furthermore, IMU data will be matched with video data to improve and explain results visually. Sensors will be placed on the RIGHT lower and upper arm of the subject (same side of the pressure mats). An additional chest mounted IMU will serve as a reference for arm (shoulder) rotation data. Three sensors will be placed on the inside of the ADS (shoulder. middle and wrist) on the RIGHT arm, positioned insider "spacer areas" and not to interfere with pilot movement or pressure mat contact.

* Commercially available APDM Opal inertial measurement unit (IMU) - 3 pilot mounted sensors: upper arm, lower arm, and chest * 3 internally hull-mounted sensors: upper and lower arm e Each is 4.8x3.6x1.3 cm (lxwxh) and weighs less than 22g

+ Lithium Ion battery at 3.7V nominal - Electrical capacity is 45OmAh. Assuming that it can last minimum 8h the current is 56 mA. With 16 hours of operation, the current is 28mA " The worst case scenario is venting of the battery. Safety precautions include aluminum base to protect the person, battery protection circuit, safe charging features. Probability estimated at .000001

The Opal movement monitor Figure 4: Opal IMU sensor from APDM

4 Detailed Test Plan This test will be performed with as many subjects as time permits. The following tasks will be performed while suited in the ADS. Each task will be repeated through 5 repetitions in 4 groups for a total of 20 repetitions. After each movement is performed, the subject will rest 2 minutes and subjective data will be collected. The subject will then repeat the movement sequence and rest period four additional times.

Below are the tasks the subject will be performing in this test campaign.

Elbow Rotation

87 I One arm executes the movement. The subject stands at the bottom of the pool. Beginning with both arms in a natural ADS position, palms facing the floor, the subject rotates at the elbow through their maximum range of motion - up and down. The subject then releases to a relaxed position.

Arm raise

The subject stands at the bottom of the pool. Beginning with both arms in a natural ADS position, palm facing the floor, the subject rotates the arms at the shoulder joint through their maximum vertical range of motion - up and down. The subject then releases to a relaxed position.

Functional Movement - Ratchet Task

The subject stands at the bottom of the pool. The subject takes the ratchet and, however must natural, seats the nut from the top of the bolt to the bottom with the use of the ratchet.

Table 1 shows a summary of all functional tasks each run will consist of, and a cumulative time for the run as currently scheduled. Each subject will perform these tasks in the order specified.

Minute Task Type Description Min Count

Elbow Rotation Isolated Stand at pool bottom and bend elbow 1.30 1.30

Arm Raise Isolated Stand at pool bottom and rotate elbow 1.30 3

Rest Rest. Qualitative Information collected 2 5

Table 1: Test variables matrix

The order of these tasks will vary between subjects. In addition, some familiarization time is built into the test plan for each suit to allow the subject to become comfortable performing each task in the suit he or she has just donned. Not only will this make the subject more comfortable and safe while performing the tasks, but it will also reduce the possibility of familiarization of a task negatively affecting the outcome of the test. (MATLAB random number generator)

This test may be terminated by the subject or test conductor at any time for any reason due to safety or hardware concerns or concern for the suited subject. Between movement groups, subjective data will be taken from the subject. This will be used as an indicator of subject fatigue and desire to terminate the test. An outline of the questions to be asked is shown in Appendix A.

The test will also be terminated in the event of ADS system failure. Standard Phoenix operating/casualty procedures will be followed regarding the failure or emergency.

88 5 Procedures

5.1 Test-Specific Pre-Test Safety Briefing

1. Anyone can stop this test at any time for any reason

2. Test personnel: Manage video camera, extension cords and functional task props at all times.

3. Suited Subject: We will ask you how you're feeling between each task, absent any other reports from you. After each series of tasks, which will last approximately 3 minutes, you will rest for 2 min.

5.2 Detailed Test Procedure

1. Initial IMU calibration

2. _ _Setup video recording system

3. Review summary of test with subject

4. Conduct test-specific pre-test safety briefing

5. __ Synchronization process

a. Turn on and detach IMUs (mark time)

6. Test personnel places pressure sensing system on subject

7. Turn on Novel (mark time)

8. Strike Extra IMU to pressure mat (mark time)

9. Test personnel places IMUs in ADS (RIGHT arm) and measures location

10. Test personnel places IMUs on the subject and measures location on the body (RIGHT arm)

11. Cover sleeve is donned

12. Body marks are pressed (1- center lower biceps, 2-center upper forearm)

Unsuited Calibration

1. Subject perform wrist pronation/supination 90 degrees (2 times)

2. Subject perform elbow flexion/extension 90 degrees (2 times)

89 3. Subject perform shoulder rotation 90 degrees (2 times)

Unsuited Familiarization Session

4. Subject practices elbow rotation

5. Subject practices arm raise

Suited Tests

1. Subject dons suit

2. Diving Supervisor performs safety checks and coordinates ADS deployment to pool bottom.

Suited Venting Pressure Test

NOTE: The subject completes 5 repetitions of the task. Time is not limited, but the task may be terminated if the subject is unable to complete 5 repetitions.

NOTE: Instruct subject to complete these tasks at what they consider to be a natural pace

NOTE: Request a report of any symptoms from suited subject after each task and ask qualitative questions from Appendix B during rest periods.

NOTE: Rest periods will be 2 minutes to assist in ease of data analysis in addition to mitigating pilot fatigue.

NOTE: Subject task order is counterbalanced for each subject and each movement run. The task order is provided in Appendix A.

NOTE: Subject may abort dive at any point

Suited Familiarization Session

1. Subject practices elbow rotation

2. Subject practices arm raise

3. Subject practices function task

Suited Data Collection Run

4. Subject performs 1st movement group

a. Allow subject to rest while prompting for subjective feedback

5. Subject performs 2nd movement group

90 a. Allow subject to rest while prompting for subjective feedback

6. _ Subject performs 3rd movement group

a. Allow subject to rest while prompting for subjective feedback

7. _ Subject performs 4th movement group

a. Allow subject to rest while prompting for subjective feedback

8. _ Subject performs Functional Task

a. Allow subject to rest while prompting for subjective feedback

9. _ Diving Supervisor coordinates ADS recovery

Post-Test Procedures

10. Suit doffed

11. IMUs removed from suit

12. _ Sync/Strike IMU to pressure pad

13. Pressure mat and IMU locations are noted for movement

14. _ Body marks surrounding pressure areas are documented

15. _ Plug in Novel Start/Stop button and press to STOP

16. _ Subject debrief (any final subjective feedback)

91

Bibliography

Allison Anderson. Understanding Human-Space Suit Interaction to Prevent Injury During Extravehicular Activity. Phd in aerospace biomedical engineering, Mas- sachusetts Institute of Technology, 2014.

T. Ciesielski and J-P. Imbert. Hydrogen Offshore Diving to a Depth of 530 m: Hydra VIII. Proceedings of Offshore Technology Con- ference, pages 289-294, 1989. doi: 10.4043/6073-MS. URL https://www.onepetro.org/conference-paper/OTC-6073-MS.

James Colgary. An Experimental Study of the One Atmosphere Diving Suit (ADS) and Data Analysis of Military Diving. Naval engineer's degree and master of science in mechanical engineering, Massachussets Institute of Technology, 2016.

Robert H Davis. Deep Diving and Submarine Operations. Saint Catherine Press Ltd, London, 6 edition, 1955.

Ryan de Vera. Submarine Recue (URC). Undersea Warfare, page Issue 56, 2014. URL http: //www. public.navy.mil/subf or/underseawarf aremagazine.

Garry L Harris. Ironsuit: The History of the Atmospheric Diving Suit. Best Publish- ing Company, Flagstaff, 1994. ISBN 0-941332-25-x.

Alexandra Hilbert. Human-Spacesuit Interaction: UnderstandingAstronaut Shoulder Injury. Master of science in aeronautics and astronautics, 2015.

Edward V. Lewis. Principles of Naval Architecture (Second Revision), Volume III - Motions in Waves and Controllability. Society of Naval Architects and Marine Engineers (SNAME), 1989.

Tony Loftas. JIM: Homo Aquatico-Metalicum. New Scientist, 58(7 June):621-623, 1973.

NAVSEA. P9290 - System for Certification Procedures and Criteria Manual for Deep Submergence Systems, Rev A. (November), 1998.

NAVSEA. US Navy Diving Manual Rev. 6, 2011.

Novel. Pliance Sensing System: Quality in Measurement. Technical Data and Use Manual. 2013.

93 David Scott. Seventy Fathoms Deep. Henry Holt and Company, New York, 1932. Steve Sullivan. Light-weight One Atmosphere ( 1 ATM ) Diving Suit. (Navy STTR FY2013A - Topic N13A-T029), 2013.

A H Techet. 2.016 Hydrodynamics, Reading #6, 2005. URL ocw.mit.edu.

Michael Albert Thornton. A Survey and Engineering Design of Atmospheric Diving Suits. Master of engineering, Texas A&M University, 2000.

Michael Triantafyllou. Design Principles of Ocean Vehicles, Lecture Notes, 2014.

94