An Experimental Study of the One Atmosphere Diving Suit (ADS) and Data Analysis of Military Diving MASSACHUSCITS INSTITUTE byI OF TECHNOLOGY James J. Colgary, Jr. JUN 0 22016 B.S., United States Naval Academy (2005)

M.S., Naval Postgraduate School (2006) - LIBRARIES Submitted to the Department of Mechanical Engineering ARCHIVES in partial fulfillment of the requirements for the degrees of Naval Engineer's Degree and Master of Science in Mechanical Engineering at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY June 2016 @ Massachusetts Institute of Technology 2016. All rights reserved.

Signature redacted Author ...... ------Department of Mechanical Engineering May,6, 2016 Certified by ...... Sig atu reedacted Alexandra H. Techet Associate Professor of Mechanical Engineering Jq1ie

Submitted to the Department of Mechanical Engineering on May 6, 2016, in partial fulfillment of the requirements for the degrees of Naval Engineer's Degree and Master of Science in Mechanical Engineering

Abstract

The Atmospheric Diving Suit (ADS) is a one-man submarine with moveable, human- like appendages with internal pressure maintained at one atmosphere. This precludes the possibility of common diving related illnesses while giving the operator an in- creased depth of operation compared to traditional diving systems. The ADS provides additional capability for industries and militaries around the world, but is not without its own unique challenges and limitations. Current ADS maneuverability, specifically that associated with joint rotation, lacks natural movement and range of motion, ren- dering most normal underwater tasks more challenging and taxing on the operator. Concerns about the lack of maneuverability and usability of the current ADS, primar- ily raised by the US Navy and ADS operators, prompted the Office of Naval Research (ONR) to fund an investigation into the next-generation ADS. In partnership under a Small Business Technology Transfer (STTR) contract, Mid6 Technology and MIT teamed up to investigate new joint design. To better understand the existing ADS and characterize the kinematics of elbow and shoulder rotation, an experimental test was completed with the commercial OceanWorks 1200 ft HARDSUITTM ADS at Phoenix International. Using a suite of Inertial Measurement Units (IMUs), equivalent ADS elbow and shoulder flexion/extension angles were extracted. A custom MATLAB® script was written to process data based on previous MIT IMU research associated with spacesuit design and other biomedical IMU research. The ADS pilot's movement inside the suit characterized the current suit's maneuverability, baselining capability. This study will inform future joint design by improving the understanding of the current ADS. In conjunction with the kinematic study, a numerical analysis of all military diving data was completed to better understand "how" the military dives. All military dive data is available to the public via www.militarydivingdata.com or divingresearch.scripts.mit.edu/militarydivingdata.

3 Thesis Supervisor: Alexandra H. Techet Title: Associate Professor of Mechanical Engineering

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

4 Acknowledgments

The following organizations, businesses, and individuals were significant to the success of this study:

a Phoenix International - Torn Bissett and the team of ADS pilots in Bayou Vista, LA.

" OceanWorks International - providing feedback and additional information on the HARDSUIT M ADS.

" Undersea Rescue Command (URC) Engineering Officer - providing ADS 2000 information and hands-on exposure.

" APDM customer support and engineers - supporting requests for information.

" Naval Safety Center - for honoring the large Freedom of Information Act diving and diving related mishap data request.

" MIT Man-Vehicle Laboratory (MVL) - for equipment support and consult.

Specifically, to Pierre Bertrand and Eddie Obropta for technical and experi-

mental support.

" Primary editor - Katie Colgary, Harvard MBA 2016, for providing critical, con-

stant, thoughtful, and loving support.

5 THIS PAGE INTENTIONALLY LEFT BLANK

6 Contents

1 Introduction 15 1.1 Motivation and Background .. . 15 1.2 Problem Statement ...... 16 1.3 Research Objectives ...... 16 1.4 Thesis Outline ...... 17

2 ADS Industry and Literature Review 19 2.1 Present Day Atmospheric Diving ...... 19 2.1.1 OceanWorks HARDSUITTM ADS .... . 20 2.1.2 Nuytco's EXOSUIT ADS ...... 22 2.1.3 ADS Operators ...... 23 2.1.4 ADS Problems and Concerns ...... 24 2.2 Kinem atics ...... 29 2.2.1 Types of Limb Kinematic Evaluation . . 29 2.2.2 Inertial Measurement Units in Kinematic Evaluation 30 2.2.3 Inertial Measurement Units ...... 30 2.2.4 Fusion Algorithms ...... 32

3 Study of ADS Arm Kinematics 33 3.1 Overview ...... 33 3.2 Experimental Setup ...... 33 3.2.1 Platform ...... 33 3.2.2 General Movements ...... 34

7 3.2.3 Location /Facility ...... 35 3.2.4 Test Subjects ...... 35 3.3 Sensors...... 36 3.3.1 Sensor Calibration ...... 37 3.3.2 Sensor placement and Orientation .. . 37 3.4 Test Protocol ...... 39 3.5 Kinematic Analysis ...... 44 3.5.1 Detailed Kinematic Analysis Methods. 45 3.6 Graphical and Numerical Analysis ...... 48 3.6.1 Elbow Rotation ...... 48 3.6.2 Arm Raise ...... 56 3.6.3 Pressure Sensor Correlation ...... 57

4 Military Diving Data 61 4.1 O verview ...... 61 4.2 Background ...... 61 4.3 Quantitative Methodology ...... 62 4.3.1 DJRS Dive logs ...... 62 4.3.2 Diving related Mishap Reports .. . . 64 4.3.3 ADS Diving Logs ...... 64 4.4 All Military Diving Breakdown ...... 65 4.4.1 Unacceptable Dive Logs ...... 74 4.5 Diving Related Mishaps ...... 74

5 Conclusions 81 5.1 Future Work ...... 83

A ADS Test Plan 85

8 List of Figures

2-1 OceanWorks current HARDSUIT Quantum. (photo credit: Ocean- Works International) ...... 20 2-2 ADS maximum elbow angle. (Photo adapted from OceanWorks Inter- national ADS HARDSUIT right arm technical drawing.) ...... 21 2-3 Nuytco EXOSUIT. (photo credit: Nuytco Research) ...... 23 2-4 Comparison of displaced suit volume and neutral buoyancy. Vertical columns show average or estimated displacement ranges. The angled line represents a minimum weight value for a given displaced volume under which, additional weight would be needed to achieve desired neutral buoyancy...... 28

2-5 APDM recommended placement of IMUs (image source: APDM) .. 31

3-1 1200 ft ADS owned and operated by Phoenix International ...... 34 3-2 Test pool facility. The ADS is being craned into position with onboard ADS camera view of pool in the upper left corner. Blue framed viewing window with externally mounted camera seen centered on pool cylinder. 35 3-3 APDM IMU Opal actual location on subject. Figure adapted from Figure 2-5...... 38 3-4 APDM IMU Opal local coordinate axes...... 38 3-5 APDM IMU Opal actual location on ADS hull...... 39 3-6 Test subject sensor placement. The blue pad is the pressure pad and the IMUs are the black-strapped objects above and below the elbow. White athletic tape helps secure sensors to the body...... 40

9 3-7 Test subject entering ADS: stepping first into the legs of the suit, then allowing the torso portion of the suit to be closed overhead and locked. 41

3-8 ADS arm rotation to achieve maximum elbow angle. (photo adapted from OceanWorks International ADS HARDSUIT right arm technical draw ing) ...... 42

3-9 Elbow rotation as captured from ADS pool testing. The central picture illustrates maximum elbow rotation ...... 42

3-10 Arm raise as captured from ADS pool testing. The central picture illustrates maximum arm raise ...... 43

3-11 Test subject performing the Functional Task: tightening a bolt with fam iliar ratchet tool...... 43

3-12 IMU numbered organization. IMUs 2 and 3 are used to extract elbow angle, where IMUs 1 and 3 are used for the shoulder angle...... 46

3-13 Typical elbow rotation output of a movement group with five repeti- tions. The difference between the peaks and valleys represent maxi-

mumn eiow imovenen...... 4

3-14 Pilot arm wide range of flexion/extension angle to achieve the same maximum ADS arm rotation. The solid lines represent pre-rotation arm orientation where the dotted lines show post-rotation orientation. 51

3-15 Variability of movement in Test Subject 2...... 52

3-16 Pilot and ADS arm flexion/ extension angles...... 53

3-17 Elbow rotation example from each subject including yaw, roll, and pitch, 55

3-18 Typical arm raise angle output of a movement group with five repe- titions. The difference between the peaks and valleys represent maxi- mum shoulder movement...... 56

3-19 Test Subject 3 elbow rotation data for a typical movement illustrat- ing two distinct force application periods of flexion (higher force) and extension (lower force)...... 58

10 3-20 Test Subject 3 arm raise data for a typical movement illustrating two distinct force application periods of flexion (higher force) and extension (low er force)...... 58

4-1 Histogram of all military diving distribution 2008-2015 with ECDF overlay ...... 66 4-2 Group normalized histogram of all military diving color coded by job description...... 68 4-3 Group normalized histogram of all military diving color coded by job description. Training specific diving types are toggled off (grayed out) to highlight the changes in normalized distribution of non-training div- ing types...... 70 4-4 Group normalized histogram of all military diving color coded by ser- vice affiliation ...... 71 4-5 Pie chart of all military diving segmented by diving apparatus type. . 72 4-6 Group normalized histogram of all military diving color coded by diving apparatus type...... 72 4-7 Histogram of all military diving segmented by year...... 73 4-8 Histogram of Navy and Marine Corps diving mishaps 2008-mid 2015 with related diving casualty type...... 76 4-9 Navy and Marine Corps diving distribution with Mishap Rate 2008-2015. 78 4-10 Navy and Marine Corps diving mishaps by apparatus type...... 80

5-1 Concept of full IMU body awareness as a means of communication for ADS diving supervisors shipboard. Note the skeleton IMU rendering on the center console representing pilot orientation and limb position. 84

11 THIS PAGE INTENTIONALLY LEFT BLANK

12 List of Tables

2.1 US Navy ADS 2000 JLU data for all dives 2006-2012. JLU is seen %11 of the tim e...... 25 2.2 APDM Opal Specifications (APDM, 2016b) ...... 31

3.1 ADS pool testing subjects. Test subjects represent a diverse mix of A D S pilots...... 36 3.2 ADS pool testing task matrix. Arm Raises are highlighted in green, El- bow Rotations highlighted in orange, and Functional Task (FT) high- lighted in purple. Each subject rests and provides verbal subjective feedback between tasks...... 44 3.3 Elbow rotation averages over three test subjects. The "--" indicates inconsistent and untrustworthy data and was not included in analysis. 50 3.4 Elbow rotation averages for pilot and hull for three tested subjects. The "--" indicates inconsistent and untrustworthy data and was not included in analysis...... 54 3.5 Arm raise averages over three test subjects...... 57

13 List of Abbreviations

ADS Atmospheric Diving Suit/System AGE Arterial Gas Embolism API Application Programming Interface COUHES Committee on the Use of Humans as Experimental Subjects DCS Decompression Sickness DJRS Dive Jump Reporting System ECDF Empirical Cumulative Distribution Function EMU Extravehicular Mobility Unit EOD Explosive Ordnance Disposal FOIA Freedom of Information Act FT Functional Task IMU Inertial Measurement Unit ISB International Society of Biomechanics JLU Joint lock-up LARS Launch and Recovery System MIT Massachusetts Institute of Technology NEDU Navy Experimental Dive Unit NWU North West Up ONR Office of Naval Research POIS Pulmonary Over-Inflation Syndrome ROM Range of motion UUA Self Contained Underwater Breathing Apparatus STTR Small Business Technology Transfer URC Undersea Rescue Command

14 Chapter 1

Introduction

1.1 Motivation and Background

The Atmospheric Diving Suit (ADS) is a one-man submarine with moveable, human- like appendages. As in a submarine, the pressure inside the pressure hull is maintained at one atmosphere throughout a diving operation. This precludes the possibility of common diving related illnesses and gives the operator, known as a "pilot," an in- creased operating depth compared to traditional diving systems. The ADS provides another facet of capability for commercial industries and militaries around the world, but is not without its own set of unique challenges and limitations. The total num- ber of active ADSs in operation is very small. With the high cost of procurement, sensitive maintenance, and required high-skill operation, many potential customers can not afford the investment, yet would benefit greatly from ADS capability. Ad- ditionally, concerns about the lack of maneuverability and usability of the current ADS, primarily raised by the US Navy, have indicated the need for an overall re- design of existing technology. Recognizing the high long term cost and often diffi- cult operability, the Office of Naval Research (ONR) is funding an investigation of a "next-generation, lightweight ADS." In partnership under a Small Business Technol- ogy Transfer (STTR) contract N13A-T029, Mide Technology of Medford, MA and the Massachusetts Institute of Technology (MIT) are working in tandem on new joint design concept development. In parallel, ONR is funding a related effort to investigate

15 novel prehensor (underwater-hand) technology under STTR N13A-TO10. The MIT research team is focused on baselining ADS performance as well as pro- viding experience and assessment of the existing ADS and military diving operations.

1.2 Problem Statement

The existing ADS has not been quantitatively analyzed for operability. This study includes experimental tests used to determine how the human pilot interacts with and maneuvers the ADS. The results of the analysis will be used to inform the new elbow joint design being developed under the STTR by Mid6. The current ADS maneuverability, specifically that associated with limb manipu- lation, lacks natural movement and range of motion (ROM), rendering most normal underwater tasks more challenging and fatiguing. Overall, ADS maneuverability must be improved if it is expected to be an asset to the industries and militaries of the fu- ture. The first step in making an ADS improvement is to understand the capabilities and limitations of the existing system.

1.3 Research Objectives

In conjunction with the focused development of an elbow joint by Mid6, this study aims to provide quantitative assessment of the existing ADS arm maneuverability. Specifically:

* Demonstrate the effectiveness of limb kinematic assessments using Inertial Mea- surement Units (IMUs).

" Establish ADS pilot ROM for elbow and shoulder flexion/extension movements.

" Provide a framework for follow-on new joint assessment using IMUs.

* Strengthen project team knowledge of military diving operations with compre- hensive quantitative assessment of diving operations.

16 The ADS experimental effort is conducted in conjunction with Naval Engineer's De- gree Candidate, Christopher Wilkins (2016). Wilkins simultaneously analyzes limb movement with the use of a pressure sensing system to numerically assess force. There will be sections of overlap between the individual studies and, in some cases, jointly generated figures will be found in both theses. Furthermore, data is exchanged to allow for a more holistic assessment of joint motion. It is recommended that both studies be referenced when researching ADS maneuverability.

1.4 Thesis Outline

This study is outlined as follows:

* Chapter 2 describes current ADSs found in operation as well as a literature review of kinematic evaluation using IMUs. A specific assessment of ADS weight limitations is also discussed in this chapter.

" Chapter 3 describes the ADS experiment and presents arm ROM data. Addi- tionally, force data from Wilkins is coupled with IMU data to reveal a more comprehensive view of kinematic motion.

" Chapter 4 presents a quantitative assessment of military diving from 2008-2015 using data acquired from the Naval Safety Center via a Freedom of Information Act request.

" Chapter 5 provides conclusions and follow-on study recommendations.

17 THIS PAGE INTENTIONALLY LEFT BLANK

18 Chapter 2

ADS Industry and Literature Review

This chapter highlights relevant research and literature related to the ADS and kine- matic evaluation using IMUs. First, a brief discussion on the current ADS is pre- sented. Although researched for this experiment, a detailed historical description of ADSs is not included but can be found in Christopher Wilkins' 2016 thesis and published work completed by Thornton et al. (2001). Highlighted in this report are current ADS uses, specific Navy ADS joint issues, and a discussion on the desire for a next-generation ADS, to include weight and depth limitations. In Section 2.2, a description of kinematic limb evaluation is also presented.

2.1 Present Day Atmospheric Diving

The ADS is a valuable asset for conducting deep-sea work and inspections when traditional diving techniques (SCUBA, Surface Supplied, Mixed Gas, or Saturation) and existing robotic systems lack the required capabilities. Primarily developed for and subsequently used as an asset to ocean-drilling oil and gas companies, the ADS has a small niche corner in the commercial undersea market (Kesling, 2013). Often called upon for repairs related to oil drilling platforms, diving contractors like Phoenix International maintain an active inventory of ADS and pilots poised for underwater construction. The ADS is also used by the scientific community, giving researchers the ability to personally experience, study, and explore the deep ocean environment.

19 From the famous deep water explorations of Sylvia Earle (1979) and Joe Maclnnis (1989) to the more modern archaeological inspection of Antikythera (2014), the ADS gives researchers unparalleled deepwater access (Kristof and MacInnis, 1983; Baehr, 2014; Kesling, 2013). Finally, the US Navy, and other militaries around the world, including France, Italy, Turkey, and Russia, maintain small inventories of ADSs for use primarily in assisting submarine rescue efforts. Reportedly, military ADSs have only been operated in training exercises. Two primary companies dominate the present day ADS design and development market: OceanWorks International (Burnaby BC, Canada) and Nuytco Research Ltd. (North Vancouver BC, Canada).

2.1.1 OceanWorks HARDSUITTM ADS

OceanWorks currently offers two HARDSUIT" models: 1200 ft and 2000 ft maximum operating depth variants. The HARDSUIT ADS is pictured in Figure 2-1.

Figure 2-1: OceanWorks current HARDSUIT Quantum. (photo credit: OceanWorks International)

As seen in Figure 2-1 the ADS consists of 20 rotary type joints capable of adjust- ing to a pilot roughly 5 ft to 6.5 ft tall. The suits can be adjusted with spacers to

provide a more custom fit to an individual pilot. All joints are individually capable

of rotating 360 degrees, with no installed mechanical stops. The human pilot inside

20 the suit activates rotation of the joint through trained limb motions. OceanWorks spokesperson Derek White estimates HARDSUIT pilots "get about 70 percent mo- bility as compared to standard diving" (Rudick, 2012). As the elbow angle is one of the target assessments in this study, Figure 2-2 illustrates maximum achievable elbow angle based on rotary joint relation and movement.

Figure 2-2: ADS maximum elbow angle. (Photo adapted from OceanWorks Interna- tional ADS HARDSUIT right arm technical drawing.)

Figure 2-2 illustrates a scale drawing of the arm as provided by OceanWorks.

Using the design software Rhino 3D, the lower portion of the arm is copied, flipped,

and superimposed on top of the existing joint (red lines). The center of mass of the

hand-pod is identified using object tools and an angle is extrapolated between the

before and after hand-pod location relative to the center of the middle-arm rotary

joint. Based on visual inspection and consideration of mechanical rotation, 81 degrees

is considered to be the maximum suit elbow flexion. Considering average humans have

21 an active flexion range of 145 degrees, the suit allows 56% elbow flexion (Lopez-Nava and Munoz-Melendez, 2015). Made out of cast (1200 ft variant) or forged (2000 ft variant) aluminum, the ADS weighs approximately 1000 lbs. In an OceanWorks provided buoyancy test report, the HARDSUIT is targeted to be slightly negatively buoyant, roughly 10 lbs, including the pilot. Without a pilot, the HARDSUIT is positively buoyant. In general, when considering total suit and pilot weight in the water, neutral buoyancy is the target value, as discussed in Section 2.1.4. The ADS is maneuvered by the pilot's feet. Contrary to popular first impressions, the pilot sits on a seat similar to that of a bicycle, rather than standing. The act of piloting the ADS has been described as "like riding a unicycle" (Baehr, 2014). The left and right feet pronate and supinate to control vertical and lateral movement through the activation of four thrusters: two vertical and two horizontal. Given the pilot is seated, the feet are free to move and manipulate the controls. The last important aspect of the HARDSUIT ADS is the life support system and tether to the surface. Equipped with a chemical resin CO 2 scrubber and recirculation fa'n th nilnt maintains Pnrhugrh replaeml nt 02 nd resin fn cprt%-f AR iveAsn wt-h 30-40 hours of reserve (OceanWorks, 2016). The pilot self-monitors air quality and adjusts an automatic 02 bellows system to affect gas partial pressures. The tether acts as the ADS's lifeline to the surface. The tether provides data transfer (voice, video, ADS depth and air quality, etc.), hydraulics, and lifting capability. The ADS is only ever manned without the tether as a result of a casualty where the tether has been severed. The tether is critical to the success of ADS deployment; given its size and weight, crane operation is necessary for launch and recovery.

2.1.2 Nuytco's EXOSUIT ADS

Nuytco's EXOSUIT ADS represents the latest in ADS technology. With many of the same features as OceanWorks' HARDSUIT, the EXOSUIT is distinguished by a lighter weight and shallower maximum operating depth. At 500-600 lbs, depending on configuration, the EXOSUIT is capable of reaching 1000 ft in depth. Although not

22 quantified, Nuytco claims "exceptional dexterity and flexibility to perform delicate work" (Nuytten, 2016). The EXOSUIT retains the ability to be configured as a swimmable suit but lacks evidence and production beyond concept design supporting the capability.

=mom AT Ole W I /* N IK7_ siz:

A

14

Figure 2-3: Nuytco EXOSUIT. (photo credit: Nuytco Research)

Overall, Nuytco is pushing the boundaries of ADS design, but neither OceanWorks nor Nuytco have an engineering solution to allow truly human-like motions with more natural flexion/ extension.

2.1.3 ADS Operators

Currently, a handful of commercial operators and five navies from around the world own and operate an ADS. Phoenix International, an underwater construction and salvage company, is one of a few commercial owners and operators of the OceanWorks ADS HARDSUIT, 1200 ft variant. Due to Phoenix International's relationship with this project's STTR, unprecedented access to their ADS and pilots facilitates the experimentation outlined in Chapter 3. JF White Contracting Co., a multi-discipline construction company, owned and operated an EXOSUIT but, in 2015, sold back the

23 system to Nuytco as a result of company restructuring. A UAE-based underwater solutions company, DeepSea Offshore International, also uses the EXOSUIT. Of the five navies that own an ADS, the French and Turkish Navies have one, the Italian Navy has three, the Russian Navy has eight, and the US Navy has four. The US Navy's ADS is a specially designed suit rated for 2000 ft, referred to as, "ADS 2000." Meant to withstand the rigors of US Navy manned submersible certification with expanded depth capability, the US Navy ADS, operated by the Undersea Rescue Command (URC), is a critical aspect of their submarine rescue plan. The US Navy, however, is planning to decommission its units in exchange for different undersea technology, and may elect to procure an unmanned system if no other improved, more viable, manned-solution is identified. As a note, oceanworks.com company profile indicates that OceanWorks maintains a relationship with four additional navies (Australia, Korea, Singapore, and Canada), presumably under a contractor or rental agreement.

2.1.4 ADS Problems and Concerns

ADS Maneuverability

More so than the commercial users of the ADS, the Navy's URC reports repeated difficulty in usability of their ADS 2000. Numerous interviews with past and present operators and those that associate with the system highlight the physical toll piloting takes on the human body, including: bruising, accelerated fatigue, and joint inop- erability (Navy ADS pilots at URC, personal communication, April 2, 2015). It is important to emphasize that no indication of long term injury is publicly reported or investigated. Specifically logged and quantified by URC, a condition known as "joint lock-up" (JLU) occurs in greater than 1 in 10 dives. This condition of JLU is defined as a time when the joint sticks or is reduced in capability as assessed by the pilot, either by reported oil leak, stiffness, or failsafe (complete joint lock). The div- ing operation may need to be aborted, losing valuable time and money, and possibly resulting in life-critical mission failure.

24 The data presented in Table 2.1 illustrates the frequency of JLU. The data, pro- vided as a courtesy from Portsmouth Naval Shipyard, partner with URC in ADS maintenance and record keeping, is generalized over all reported joint issues.

Table 2.1: US Navy ADS 2000 JLU data for all dives 2006-2012. JLU is seen %11 of the time.

ADS dives recorded from 2006-2012 Suit 1 Suit 2 Suit 3 Suit 4 Total Dives

Number of Dives 93 107 117 114 431

Number of JLU affected 6 15 16 10 47 dives

Number of individual JLUs 63 77 125 85 350

Percent JLU 6% 14% 14% 9% 11%

Avg expected number of JLUs per failure (nearest 11 5 8 9 8 whole number) __

Although the Navy's data indicates an average JLU occurrence of 11%, with a minimum of 6% and maximum of 14%, Navy ADS pilot interviews indicate the joints are often very hard to move. This fact lends ambiguity to "stiffness" qualifying as JLU and may indicate actual JLU is more prevalent than indicated in Table 2.1. Regardless of the actual quantitative assessment of joint operability, general pilot assessment is that joint movement is "difficult and stiff" (Navy ADS pilots at URC, personal communication, April 2, 2015).

ADS Weight

Also highlighted by the US Navy is the high, often cumbersome, weight of the total system. Specifically identified as a goal in STTR N13A-T029, the Navy wishes to see a lightweight system. Although commercial systems weigh 500 (EXOSUIT) to 1000

25 (HARDSUIT ADS, 1200 ft variant) lbs, the US Navy's ADS 2000 can weigh as much as 1400 lbs. When considering the total launch, recovery, and support system that must accompany the ADS 2000, the weight footprint skyrockets to about 80,000 lbs as highlighted in Navy's Scope of Certification Notebook for the ADS 2000 (2005). The lighter weight EXOSUIT system in contrast, boasts roughly 2000 lbs of support gear (MD, 2009). A physically large component of the US Navy's ADS operation is a specially designed Launch and Recovery System (LARS) that, by itself, weighs over 40,000 lbs. As a goal of the STTR study, the hope is that a lighter weight ADS brings with it less heavyweight support equipment.

Although weight may be a good metric to measure ADS design success, it is crucial to consider the lower weight limit inherent with underwater operation. Neutral buoyancy, the condition of not floating nor sinking, is a mathematically predictable condition. Being too negatively buoyant results in a high rate of sinking, and being too positively buoyant results in bobbing on the surface, potentially causing uncorrectable instability. When designing submersibles, neutral buoyancy is a desirable goal and subsequently defines a lower weight limit to ADS design.

In the case of the ADS, required submersible volume is a distinct design driver. The ADS must allow a person of average size to be enclosed inside the suit. The ma- terial that surrounds the ADS pilot will have some amount of thickness, proportional to material type and maximum operating depth, as well as a separation distance between the pilot's body and the inner material diameter. In the torso region, sepa- ration distance is normally increased to allow a pilot to extract their hands from the arms during launch, recovery, to manipulate atmosphere control valves, or simply to scratch their face. Adding to the torso volume increase is the necessity for atmosphere control. In current ADS models, CO 2 absorbent and a recirculating fan are placed directly behind the pilot. The CO 2 scrubber does not need to be housed inside the ADS pressure hull; it could be placed external in a standard diver's rebreather con- figuration. Regardless, an increase in ADS design volume is incurred whether designs include atmosphere control inside or outside the pressure hull. Lastly, other external mounted components like thrusters, cameras, SONAR, 02 bottles, and other small

26 miscellaneous items all contribute to the ADS design volume. Using force balance methodology, the buoyant force must be equal to the force of gravity on the ADS to achieve neutral buoyancy. Equation 2.1 shows this mathemat- ical relationship.

FB = FgADS(21

Understanding that:

FB Pwater VADS g (2.2)

FgADS = MADS g (2.3)

where:

Pwater = 1027 kg/m 3, density of seawater

VADS = Volume of water displaced by the submerged ADS mADS = mass of a dry ADS

g = 9.81 m/s2 , acceleration due to gravity

Knowing that there is a weight goal of 400 lbs outlined by STTR N13A-T029, a chart helps visually illustrate the "Neutral Buoyancy Line." Figure 2-4 shows the Neutral Buoyancy Line with an average human (-170-185 lbs), spacesuit, EXOSUIT, and ADS volumes overlaid to show comparative sizing. Spacesuit volume sizes are collected from the work of Freudenrich Ph.D. (2000) for science.howstuffworks.com and references the classic Extravehicular Mobility Unit (EMU) type spacesuit. Nei- ther Nuytco nor OceanWorks maintain or directly determine buoyant volume values, therefore the volumes identified on the Figure 2-4 are determined through manipula- tion of Equations 2.2 and 2.3 based on known weight of the ADS, average diver weight (185 lbs) and assumption of neutral buoyancy. In general, as mentioned previously in Section 2.1.4 a submersible should be neutrally buoyant when submerged. However,

27 ll ------= - _7MIIIIIIIIII

in the case of the ADS, systems tend to be designed slightly negatively buoyant (~10 lbs) to ensure ease of descent. In case of emergency, the ADS pilot can jettison heavy thrusters to become positively buoyant and ascend to the surface.

Comparison of Displaced Suit Volume and Neutral Buoyancy

1400

1200

1000

A verage Human Volume

800 Average Spacesuit Volume Average ADS Volume

Exosuit Estimated Volume 600 Neutral Buoyancy Line

Proposed Weight Target

400

200 -

0 2.5 3.5 4.6 5.7 6.7 7.8 8.8 9.9 10.9 12.0 13.1 14.1 15.2 16.2 17 3 18.4 19.4 20.5 21.5 Volume (cuft)

Figure 2-4: Comparison of displaced suit volume and neutral buoyancy. Vertical columns show average or estimated displacement ranges. The angled line represents a minimum weight value for a given displaced volume under which, additional weight would be needed to achieve desired neutral buoyancy.

Figure 2-4 can act as a first pass tool for ADS designers and acquisition decision-

makers. For example, if designers take the current volumetric envelope of the Ocean- Works ADS and use a significantly lighter composite material to construct the hull, ultimately, the system would not change weight. This is due to the fact that lead, or some other dense material, must be added to the system prior to entering the water. Otherwise, the ADS would be too buoyant and likely not submerge. Overall, the Neutral Buoyancy Line is critical to observe in initial design stages of underwater systems.

28 Finally, the proposed weight target for the ADS is illustrated on Figure 2-4 as a red line. This line has a value which takes the goal value listed in STTR N13A-T029 and accounts for the pilot weight (400 + 185 = 585 lbs). Therefore, one can see that the STTR targeted weight goal is a difficult task and requires a reduction in displaced volume even over that of the advanced, low volume, EXOSUIT.

2.2 Kinematics

Detailed quantitative assessment of limb kinematics is a critical step in a human suit design process. Studies in spacesuit kinematics over the past 50 years (1966-2016) contribute to increasing astronaut performance, as well as minimizing astronaut in- jury and energy expenditure (Bertrand, 2016). Specifics about suit maneuverability help inform and improve future spacesuit designs. The medical research community also widely participates in limb kinematic research, striving to assess patient limb mobility using precise methods to characterize ROM. Advanced measurement tech- nologies to support kinematic evaluation are researched and widely implemented in the development of the International Society of Biomechanics (ISB) best practices to support the use of advanced technology in medical ROM assessments.

2.2.1 Types of Limb Kinematic Evaluation

Kinematic evaluation of spacesuits is possible using a variety of techniques. Given the close similarities in desired output to the spacesuit evaluation, this study aims to use prior spacesuit research as a baseline to experimentation. Unmanned robotic testing, photogrammetry, optical video capture, and IMUs all act as effective tools for anal- ysis. Unmanned spacesuit testing using robotic actuation leads to good quantitative kinematic data but is often founded in expensive robotic infrastructure. Photogram- metry, the processes of overlaying multiple pictures to characterize single axis ROM assessment, allows for a simple yet constrained assessment. Optical motion, or video capture, the present day "gold standard" for joint angle assessment, tracks defined points on a body in three dimensions. In fact, the use of this technology is sometimes

29 employed underwater in spacesuit testing (Reinhardt, 1989). Finally, using small, inexpensive IMUs is an effective method to assess kinematic motion (Bertrand, 2016; Kobrick et al., 2012). Each method, discussed by Bertrand in his 2016 thesis, re- veals distinct advantages and disadvantages, all contributing to the evolution of the spacesuit design.

For the purposes of this study, preference is given to the use of a highly portable, inexpensive, and available tool for ADS kinematic evaluation. Therefore, a suite of IMUs is employed for experimentation.

2.2.2 Inertial Measurement Units in Kinematic Evaluation

Founded in medical joint angle assessment, IMUs perform well in revealing limb ROM. Researchers such as Favre et al. (2008), Ang et al. (2013), El-Gohary (2013), Oberlinder (2015) provide exhaustive and continued research into IMU-based limb kinematic assessment. A popular topic in the Journal of Biomechanics, this method of evaluation is constantly being assessed, compared, and employed in various manners to aid in the quantitative assessment of biomechanics.

2.2.3 Inertial Measurement Units

Inertial measurement units are small electronic devices normally consisting of one or more accelerometers and gyroscopes. Additionally, with more modern IMUs, a magnetometer is included to help calibrate against accrued bias or drift of the ac- celerometers and gyroscopes. Often three orthogonally mounted accelerometers and gyroscopes are found within an IMU to measure three-dimensional accelerations and rotations, respectively (Oberlknder, 2015). Oberldnder completes an exhaustive de- scription of IMU functionality in his 2015 report highlighting details of individual accelerometer and gyroscope construction and operation.

30 Selected Inertial Measurement Units: APDM Opal

Primarily based on availability, the selected IMU for this study is the APDM Opal. The MIT Man-Vehicle Lab has used the Opal for various spacesuit and human subject tests in the past and provides the testing equipment for this study. Although a rela- tively small company, APDM is a part of over 220 published studies with more than 25,000 test subjects, lending credibility to the product (APDM, 2016a). Table 2.2 shows the specifications of the Opal IMU.

Table 2.2: APDM Opal Specifications (APDM, 2016b)

Accelerometer Gyroscope Magnetometer

Axes 3 3 3 Range 2g or 6g 2000deg/s 6 gauss Output Rate 20-128 Hz 20-128 Hz 20-128 Hz

Additionally, Figure 2-5 from the APDM website shows the relative size and generic body placement of the IMU during full-body kinematic assessment. Chapter 3 includes graphics to indicate IMU placement in this study.

Figure 2-5: APDM recommended placement of IMUs (image source: APDM)

31 APDM Opal is highly accurate when compared to the VICON optical system, a regarded "gold standard" (Bertrand, 2016). Furthermore, Bertrand identifies mea- surement consistency through temperature changes, large rotations, and assesses low static drift over long duration readings. For these additional reasons, the Opal is an ideal IMU for ADS kinematic research.

2.2.4 Fusion Algorithms

A key facet of IMU implementation is combining inputs from the accelerometers, gyroscopes, and magnetometer. Many papers and studies dedicate research to refining Kalman filters in order to accurately define IMU rotation and position based on inputs from the multiple components inside the IMU. This study relies on the proprietary APDM filtering algorithm based on recommendations from industry professionals and comparison results shown by Bertrand (2016). Furthermore, given the infancy of the ADS kinematic testing framework, Fusion Algorithms are not explored in depth.

32 Chapter 3

Study of ADS Arm Kinematics

3.1 Overview

This chapter discusses the experimental design, sensors employed, data analysis, and results of testing of the OceanWorks 1200 ft HARDSUIT ADS in May of 2015 at Phoenix International, Bayou Vista, LA.

3.2 Experimental Setup

As discussed in Section 1.3, this research intends to prove the effectiveness of ADS kinematic testing while aiding the development of new ADS joint technology by baselining performance of the existing ADS and providing a framework for future joint assessment. The scope of the experiments and analyses are driven by available resources, time, and funding.

3.2.1 Platform

While the goal of the study is to support the US Navy STTR N13A-T029, the com- mercial 1200 ft variant HARDSUIT, pictured in Figure 3-1, provides the platform for experimentation rather than the US Navy ADS 2000 for several reasons. The driving factor for this decision is due to the US Navy's off-gassing requirements found in the

33 NAVSEA P-9290. This document requires every item present in a non-ventilated, enclosed space (the ADS) be certified. through individual testing. as non-hazardous to the operating pilot. On the order of $30,000, the Navy certification test could not he completed within the scope of this study. Alternatively, Phocnix Intcrnational allows sensor use inside the ADS for this study based on a detailed materials list, NASA approval letter from MIT IMU and pressure pad spacesuit testing, and MIT's Conunittee on the Use of Hunians as Experimental Subjects (COUHES) written ap-

proval. Next, the STTR target ADS depth capability of 1000 ft, is more comparable to the commercial HARDSUIT depth rating and therefore a better representation of an analogous system. Finally, the joint performance on the 1200 ft HARDSUIT is superior to that of the ADS 2000 for reasons discussed in Section 2.1.4 and therefore is, provides a better baseline of current joint, capability for future comparison. That if joint performance of one of the top performing modern designs is baselined, it is ultimately more valuable in informing future joint design.

/4

Figure 3-1: 1200 ft ADS owned anld operated by Phoenix International

3.2.2 General Movements

The experiment's design allows the collection of level of effort and ROM data when the ADS pilot moves their arm through a regulated set of simple mnotions. To minimize movement complexity, the motions focus on singular rotations of an ADS joint ill

34 order to simplify data analysis and improve repeatability. Additionally, because this experiment represents the first, modern, quantitative look at the human interaction with the ADS, the simple movements allow basic determination of sensor viability and testing feasibility.

3.2.3 Location/ Facility

The experiment takes place in the Phoenix International ADS saltwater training pool in Bayou Vista, LA. The pool, seen in Figure 3-2, is a vertical cylinder with a diameter of 11 ft and a depth of 13 ft,. The test pool incorporates a large observation window accessible from the ground for visual communication and surveillance. Cameras put inside and outside the tank provide in-situ and post test observation and record. Each pilot test subject operates an appropriately sized HARDSUIT discussed in 3.2.1.

Figure 3-2: Test pool facility. The ADS is being craned into position with onboard ADS camera view of pool in the upper left corner. Blue framed viewing window with externally mounted camera seen centered on pool cylinder.

3.2.4 Test Subjects

COUHES test subject voluiiteer procedures provide the framework to request par- ticipation in this study. Prior to the experiment the test examiners provided test

35 procedures, equipment information, and correspondence to the Phoenix, Bayou Vista management team and ADS pilots. As mentioned in Section 3.2, resources (pilots) were limited, but all four available pilots volunteered and appear as test subjects in this study. Given the infancy of this test protocol, using just, four pilots allows ad- equate assessment of method, sensors, forces, and ROM but, in the future, a larger group of test subjects should he recruited and assessed to find great er statistical sig- nificance in analyses. Table 3.1 shows the characteristics of the participating ADS pilots. Note the large variation in size, weight. and experience. Overall, the test subjects represent a diverse grouping of pilots.

Table 3.1: ADS pool testing subjects. Test subjects represent a diverse mix of ADS pilots.

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" 180 10

3.3 Sensors

In order to understand the kinematic ROM and effort an ADS pilot experiences when

moving the arm, the test examiners employ a sensor system of IMUs and a Novel®

pressure pad. As discussed in Section 2.2.3, the IMU sensor, APDM Opal, records

kinematic motion for evaluation. There are six total IMUs oi the pilot (three IMUs)

and ADS hull (three IMUs) to record motions of the pilot and the ADS independently.

The single Novel pressure pad, just above the pilot's wrist, reveals related pressure

and force data. The pressure pad placement is slightly different for each pilot, but

is at the centralized location of pilot-arm" ADS-arm interaction during movement.

Wilkins' 2016 thesis provides in-depth pressure pad discussion and analyses whereas

this study focuses on IMU implementation and data.

36 0 3.3.1 Sensor Calibration

Pre-experimentation testing and IMU evaluation revealed no major calibration issues but the less intensive user-calibration is applied onsite to ensure IMU collection accu- racy. Factory calibration is assumed acceptable based on past user experience, assess- ment of strong signal stability over time, and the absence of an IMU-damaging event. If major deviations were observed, the IMUs would have been factory re-calibrated by APDM. Based on the exposure of the IMUs to air travel and large change in geo- graphic location, a user-initiated, software-based magnetometer and gyroscope "user calibration" is necessary. Using APDM's MotionStudio@ software, which commu- nicates and imports IMU data, the on-screen instructions for user calibration are followed and applied. In addition to software based IMU calibration, calibrating movements are directed to assess IMU ROM calculations. Subjects stand with arms naturally at the sides, palms facing inward. When directed, the test subject executes wrist pronation and supination (maximum: 90 degrees), elbow flexion and extension (maximum: 90 de- grees), and shoulder flexion and extension (maximum: 90 degrees). This method of dynamic calibration, found in many research studies, applies an offset to IMU data based on alignment and actual biomechanical joint movement (Favre et al., 2009; Bertrand, 2016). Given the goals of this study, however, the movements provide a visual graphical validation of sensor alignment and numerical offsets are not applied. For all test subjects, the calibrating movements show expected results. In future experiment iterations, a more in-depth dynamic calibration and careful alignment procedure should be implemented to raise accuracy and precision of results.

3.3.2 Sensor placement and Orientation

IMU positions are on the test subject's torso, lower arm, and upper arm as seen in Figure 3-3. In addition, to ensure consistency of analysis, each local IMU x-axis points down. Figure 3-4 shows the three local axes. For in situ axis recognition, the positive x-axis orientation is identifiable by the USB connection port.

37 ( U

I" / / K-i

Figure 3-3: APDM IMU Opal actual location on subject. Figure adapted from Fig- ure 2-5.

z

Y

Figure 3-4: APDM IMU Opal local coordinate axes.

Additionally, the sensors position on the ADS hull can be seen in Figure 3-5. Using a quick drying adlesive and self-adhesive Velcro pads, sensors mount to the ADS arm and remain in place for the duration of the experimentation period. The quick drying, two-part adhesive is often used by the Phoenix pilots and is reconmneinded for sturdy yet removable application. The two-part adhesive secures one side of the Velcro to the hull to ensure a reliable hold given the inside of the hull is often slick with lubricating oil from the joints. The other side of the Velcro is directly applied to the EMU with its self-adhesive. Two-part adhesive is not, used directly on the IMU for concerns of damaging the thin plastic case. Despite the concern given to firm placement, the

38 attachment method proves only marginally successful and results in some movement during subject testing, rendering roughly 40% of the data unusable.

Figure 3-5: APDM IMU Opal actual location on ADS hull.

3.4 Test Protocol

Test examiners turn on and initialize all systems (camera, pressure pad, and IMUs) according to their individual operating procedures. Given all separate systems do not intercommunicate, an application of a mechanical syncing technique ensures post processing time consistency among sensor analyses. First, throughout an entire sub- ject's test, a stopwatch records events and acts as the main information source for post-test data splicing. Additionally, while on video and prior to placement, the test examiners use the torso IMU to "tap" the pressure pad three times to produce re- sponse in both systems. During post-test processing, the three "spikes" in both the

IMU and pressure systems are aligned to identify a time offset.

After calibration and time sync, the test subject completes remaining ADS out- fitting procedures in the warehouse adjacent to the test pool. While sensors record data internally for analysis purposes, they also transmit data via Bluetooth in order

39 to verify system functionality prior to ADS suit-up. Figure 3-6 shows an example of sensor placement. The torso ILMU is misplaced in the picture and is being used for the "tap" sensor alignment discussed above.

Figure 3-6: Test subject sensor placement. The blue pad is the pressure pad and the IMUs are the black-strapped objects above and below the elbow. White athletic tape helps secure sensors to the body.

Once test subjects are inside the ADS, the Phoenix International (living sulper-

visor has operational and safety control of the ADS systems including life support,

communications, thrusters, as well as transport of the suit and pilot via crane into

the test pool, seen in Figure 3-2 above. All operations are to be conducted accord-

ing to Phoenix's own safety procedures with no interference from the experimental

e(lilpment or examiners. Figure 3-7 shows a test subject entering the ADS. Notice

the subject enters via the ADS midline allowing the top of the system to close like a

clamshell.

40 Figure 3-7: Test subject entering ADS: stepping first into the legs of the suit, then allowing the torso portion of the suit to be closed overhead and locked.

After the test subjects are in the pool and positioned near the viewing window, they familiarize themselves with the movements in the ADS to prevent an appreciable learning curve skewing performance during the experiment.

The evaluation includes two specific movements. The first is known as an "elbow rotation," focusing rotation at the fourth joint froin the shoulder. Pilots attempt to move the ADS arm through its full range of motion while working to achieve maximum ADS arm "bend." Figure 3-8 shows the 2D graphic representation of movement. While the motion is meant to represent a normal human "elbow bend," the ADS arm requires both rotational and flexion/extension to achieve the desired result. Figure 3-9 breaks down the pilot's actual movement during pool testing.

41 0 0

Figure 3-8: ADS arm rotation to achieve maximum elbow angle. (photo adapted from OceanWorks International ADS HARDSUIT right, arm technical drawing)

Figure 3-9: Elbow rotation as captured from ADS pool testing. The central picture illustrates maximum elbow rotation.

The second motion is a shoulder flexion extension focusing on the top joint seen in Figure 3-8. Known as the "arm raise," pilots keep their arm extended and move fromn a relaxed position, up, and down, raising the extended arm above their head through maximnun ROM. The arn raise is seen in Figure 3-10 and although the ADS arm is bent, the pilot's arm is extended and only rotates at the shoulder.

42 Figure 3-10: Arm raise as captured from ADS pool testing. The central picture illustrates maximum arm raise

After the completion of the ell)ow rotation and arm raise, the pilots perform a task with a focus on operational relevancy rather than ai defined imotion. Known as the "functional task," it, emulates motions actually performed by ADS pilots during a normal working dive. The functional task involves securing a flange by tightening a nut onto a bolt. Using a standard socket wrench adapted for use with a wrenching lanyard (standard practice), the ADS pilot grips the tool, places it on the project site, and tightens the nut. Although thought of as a simple movement, most subjects denonstrate more difficulty and general inconsistency than anticipated. This assess- mient ultimately becomes a qualitative data point in the study and is not numerically assessed in Section 3.6. A photograph of the modified tool and task setup in use by an ADS pilot during pool testing is shown in Figure 3-11.

Figure 3-11: Test subject perfornming the Functional Task: tightening a bolt with familiar ratchet tool.

Each motion, with the exception of the functional task, repeats 20 times per

43 test subject in four "movement groups" of five repetitions. The order of each motion group is different, for each subject through application of a random sequence. Between each movement group there is a timed rest period. In addition, some familiarization time in the test plan allows the subject to become comfortable performing each task prior to testing. Not only does this increase the test subjects comfort and safety, but also reduces the possibility of familiarization-effect s which may ultimately reduce the amount of viable data. During each rest period, pilots provide verbal subjective feedback. Feedback enables test examiners to monitor safety and is not, quantitatively processed. The testing matrix, Table 3.2, is the final testing breakdown with the order created using a simple random number generator script in IATLAB@.

Table 3.2: ADS pool testing task matrix. Armi Raises are highlighted in green. Elbow Rotations highlighted in orange, and Functional Task (FT) highlighted in purple. Each subject rests and provides verbal subjective feedback between tasks.

Task Task Task Task Task Task Task Task Task

Sbect 1 arm arm elbow elbow arm elbow arm elbow e - !uraise I raise 2 rot 1 rot 2 raise 3 rot 3 raise 4 rot 4

elbow arm elbow m arm -o arm -0 elbow arm elbow Subject 2 -0- '-101 FT rot 1 raise 1 rot 2 raise 2 raise 3 rot 3 raise 4 rot 4 U.U. UL. U.. U- LL LA-

C C C C CC I arm m elbow c arm m arm M elbow M elbow M elbow m arm F 'raise 1 rot 1 raise 2 " raise 3 rot 2 rot 3 rot 4 raise4 FT C3 - -1- C

Subject 4 arm arm elbow arm arm elbow elbow elbow FT ra ise 1 raise 2 rot 1 raise 3 raise 4 rot 2 rot 3 rot 4

The test procedure document is included as Appendix A for reference and repro-

duction.

3.5 Kinematic Analysis

IMU data is locally stored in the individual unit's onlboardi memory and after each

test, the test examiners imnport data to APDM's MotionStudio software. The soft-

ware is able to capture and display live orientation data, however, given the nature of

44 the test, live data is only viewed at the start of each test to ensure IMU functionality. MotionStudio's most important function in this study is to provide a means to import and export individual test data. The large CSV files are transferred from MotionStu- dio to MATLAB for analysis. Containing all applicable three-dimensional sensor data from the accelerometer, magnetometer, and gyroscope, the CSV files provide all the data necessary for analyses including the filtered IMU orientation in quaternion form. As discussed in Section 2.2.4, the powerful and accurate APDM algorithm provides the IMU output accuracy for this study and therefore, filtered quaternion output is used in MATLAB for analysis. Quaternions can offer great computational advantage when analyzing data. How- ever, given the unique make-up of the four coordinate, real/complex, scalar/vector combination, specific quaternion math is appropriate. Because efficient computational analysis is not a goal of this study, a more descriptive method is used to emphasize understanding and application. Quaternions are converted to rotation matrices using methods primarily outlined in the journal paper written by Kobrick et al. (2012), which allows the use of standard matrix math throughout analysis. Using rotation matrices, one IMU reference frame is transformed into another to determine a third rotation matrix which represents IMU relative movement over a time interval. Euler angles are extracted for both the elbow rotation and the arm raise described in Sec- tion 3.4 and are estimates of ROM for elbow and shoulder flexion/extension of the pilot and ADS arm.

3.5.1 Detailed Kinematic Analysis Methods

A MATLAB script first imports each IMU's dataset for a given test subject and applies IMU numbering standardization in accordance with Figure 3-12. Next, a MAT file saves common time and quaternion data (time, QuatX1, QuatYl, QuatZ1, QuatX2, QuatY2, QuatZ2, etc.) for all six IMUs in a given test. Concurrently, individual time variables for each movement group are extracted from a separate file to allow ease of movement group identification and analysis. Throughout analysis, timestamps on movement groups are updated based on IMU data visual behavior.

45 2$

Figure 3-12: IMU numbered organization. IMUs 2 and 3 are used to extract elbow angle, where IMUs 1 and 3 are used for the shoulder angle.

To begin analysis, the MATLAB function "quat2angle" takes quaternion values from a given test and converts it into useable Euler Angles (in radians). Specifically, the function executes Equation 3.1 but allows for more simple, readable, and quickly implemented methodology.

arctan 1-2(q2+q)

0 arcsin(2(qOq2 - q3qi)) (3.1) [. arctan 2 (qoq3+qq2) [<7 I -2(q22+q3)_

where:

T -T

q [ qo q1 q3 q4 ='s : y I = R oll

O = Pitch

0 = Yaw

Once each IMU (1-6) represents its orientation in angle form (Roll, Pitch, Yaw), related rotation matrices in the North West Up (NWU) common reference frame produce the starting point of the next calculation. Rotation matrices, using the

46 MATLAB function "angle2dcm," execute Equation 3.2. Note the example below shows methodology for calculating elbow angle (Wie, 2008).

c(9)c() c(0)s(4') -S(6) NWU R IMU = s(#)s()c(/) - c()s() s(#)s(6)s(/ ) + c(#)c(/) s(#)c(0) (3.2) c(#)s(O)c(') + s(#)s(Wb) c(#)s(0)s(b) - s(#)c(#) c(0)c(6)J

where:

RE i = Rotation matrix from IMU to B.

B, in this case, represents the NWU common reference frame

c = cos(angle)

s = sin(angle)

To reveal angles associated with the elbow rotation, IMU 3 must mathematically rotate to IMU 2 reference frame. To reveal angles associated with the arm raise, IMU 3 must mathematically rotate to IMU 1. The matrix chain-rule multiplication of the transpose produces the desired output. For example, Equation 3.3 shows the rotation of IMU 3 to the IMU 2 reference frame.

IMU2 NWU NWU R IMU3 = RIMu3 (RIMU2) (3.3)

Using the methods by Kobrick et al. (2012) and Slabaugh (1999), Euler angles are extracted from the rotation matrix and plotted. Equation 3.4 represents the methodology. In the case of the elbow rotation, yaw represents ROM from extension to maximum flexion.

47 if (R3 1 # 1) 01 = -asin(R 31 ) 02 = 7r - 01

V)1 = atan2 R3 2 R33 Cos 0 1 1 Cos 01 R32 R3 3 2 = atan2 COS 02 COS 02

01 = atan2 c1 9os 'o01 . 11 02 = atan2 R21 2 else q = anything; can set to 0 if (R31 = -1) 9 = r/2 = q + atan2(R , R ) else 12 13 0 -7r/2 ?= -0 + atan2(-R , -R1 ) end if 2 3 end if (3.4)

For clarity, at the start of each motion group the angle value is zeroed and plotted with a MATLAB algorithm to identify maximum and minimum values.

3.6 Graphical and Numerical Anialyi

During initial inspection of IMU data and angle results, only three of the four test subject's data is useable for presentation. In the fourth and final test, IMU data is inconsistent and unreadable. The test examiners assume it is due to a faulty cali- bration or unrecoverable import fault. All other test subjects contribute acceptable results. To further ensure the accuracy of data in test subjects 1-3, inspection of the magnetic field trace shows consistency throughout each test, ensuring the APDM fusion algorithm remains valid.

3.6.1 Elbow Rotation

Of the 80 possible elbow rotations, only 60 are analyzed. With data from the re- maining three test subjects, flexion/extension angles are extracted using the method outlined above. Figure 3-13 illustrates a typical graphical output from an elbow

48 rotation movement group.

Typical Elbow Rotation 90 V elbow angle peak 80 V U valley V 70 I 60 'I ~ / / 50 ) 1' 40

0) 30 / / I / I 20 I / / 10 / a / II

0

-10 L 2000 0 200 400 600 800 1000 1200 1400 1600 1800 Sample (128=1 sec)

Figure 3-13: Typical elbow rotation output of a movement group with five repetitions. The difference between the peaks and valleys represent maximum elbow movement.

Elbow movement is generally smooth, deliberate, and well paced. Pilots do not rush and move the ADS arm as they would under normal working circumstances with a 2-3 second period of motion. As seen in Figure 3-13 and consistently among all pilots, each individual movement does not always start with the same arm exten- sion, varying by a few degrees each period. Flexion /extension angles are a result of the magnitude of change between related peaks and valleys and not from a straight maximum determination from the zero x-axis. Table 3.3 lists pilot flexion/extension angle averages over the duration of the experiment. As discussed in Section 3.4, pilots move the ADS arm through maximum

49

I - -1W mechanical rotation.

Table 3.3: Elbow rotation averages over three test subjects. The "--" indicates in- consistent and untrustworthy data and was not included in analysis.

Mean Angle (deg) Subject 1 69.3 Subject 2 94.4 Subject 3 _ 74.7 Subject 4 Total Avg 79.4 Stdev 21.1

In many instances, it appears as though the pilots do not achieve maximum me- arm chanical rotation based on pilot arm data alone. Furthermore, based on human the movement data from inside a suit from Kobrick et al. (2012), it is expected that two pilot arm would move through a greater angle than that of the ADS arm. These

concerns prompt further investigation.

Given the wide ADS arm diameter and rotary joint mechanical motion, the pi-

lot arm can achieve maximum ADS arm rotation through a wide range of angles. the ADS Figure 3-16 illustrates the estimate ranges of pilot arm movement within pilot flex- to achieve maximum ADS arm movement. It may be possible to see while ion/extension angles from as low as ~60 degrees to as high as -135 degrees arm place- achieving the same ADS arm angle. This estimation is based on general rather, ment and does not take into account average anthropomorphic measurements; data va- it's ali illustration of the flexibility in pilot movement as an aid in assessing

lidity.

50 60

Figure 3-14: Pilot arm wide range of flexion /extension angle to achieve the same maximurn ADS arm rotation. The solid lines represent pre-rotation arm orientation where the dotted lines show post-rotation orientation.

The average pilot elbow flexion /extension angles inside the ADS in Table 3.3, plus or minus the standard deviation, show a wide range of kinematic motion and speaks to the variability of movement. Variability amongst all subjects exists. Figure 3-15 illustrates the variability in elbow rotation of Test Subject 2.

51 in Elbow Rotation 160 Test 2 Varalbillty

140

120

100

S80!

60

40

400 a

-20: 1 - A ______0 50 100 150 200 250 300 350 Sample (128=1 sec)

Figure 3-15: Variability of movement in Test Subject 2.

Test Subject 2 is generally consistent for 63% of elbow rotations. Low elbow angles during the first movement group and two very high elbow readings in the last, show significant deviation from the mean. They are possibly due to pilot internal adjustments which cannot be verified with video or post-test pilot feedback. Overall, the other test subjects exhibit similar results, including some unexplainable outliers which are not included in analysis. IMUs on the ADS arm hull as seen above in Figure 3-5, provide a reference capability on actual ADS arm rotation. It was assumed the pilot could move the ADS arm through maximum rotation (81 degrees), but often, this is not the case. Although affected by poor placement, securing method, and accidental pilot contact, the hull mounted IMUs still provide some interesting data to understand isolated movement groups, but do not yield enough consistent data for viable statistical analysis. With good data from Test Subject 2, Figure 3-16 shows an example of different pilot arm positioning while rotating the ADS arm at the elbow. In the first elbow rotation movement group, the pilot bends his arm less than the achieved ADS arm angle and in the second elbow rotation, the opposite is observable. Additionally, Figure 3-16

52 shows that even with the pilot attempting to reach maximum ADS arm movement, the maximum achievable ADS arm angle is not reached.

so Test 2 Elbow Rotation I P"10 Hull 70

60

50

40 0 2 30 02 02 C 20

10 I I 0

-10

-20 0 500 1000 1500 2000 2500 Sample (128=1 sec)

Test 2 Elbow Rotation 2 90 Not Hull 80

70

60

50

10

30 20

130

20

-10 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Sample (128=1 sec)

Figure 3-16: Pilot and ADS arm flexion/extension angles.

Table 3.4 lists pilot and hull elbow rotation angle averages over the duration of the experiment. Some data is repeated in more detail from Table 3.3 for comparison. Note

53 the "--" indicates inconsistent aln(l untrustworthy data based on trace assessment.

As noted above. there is not enough quantifiable data to make a solid correlation between pilot, and hull angle, but it is qualitatively interesting to see the variety of angles. illustrating the relative flexibility of movement inside the ADS. Additionally, zero of three pilots achieve the maximum expected ADS arm movement.

Table 3.4: Elbow rotation averages for pilot and hull for three tested subjects. The " indicates inconsistent and untrustworthy data and was not included in analysis.

Subject 1 Subject 2 Subject 3 E R Pilot (dee) I Hull (deg) Pilot (deg) Hull (deg) Pilot (deg) i Hull (deg) 1 61.59 44.47 54.84 77.59 76.27 39.92 2 73.92 -- 81.29 67.75 75.63 72.43 3 72.33 11.63 -- 72.06 70.49 4 -- 119.63 - -71.61 Mean 69.28 44.47 94.36 72.67 74.65 63.61

Figure 3-16 also illustrates the order of motion initiation. As expected, the pilot

arm moves first, the affecting the movement of the ADS arm. This concept further draws attention to the roomy ADS arm and its divergence from more tightly fitting clothing or spacesuits which generally move as an extension of the body.

Not only is elbow flexion /extension angle extracted in analysis, but also other mo- tion characteristics. Figure 3-17 shows a sample of elbow rotation from each of the test

subjects, displaying yaw (elbow flexion extension angle), roll (l.)romiation 'supination), and pitch (biornechanical link to roll).

54 internal: test 1 Elbow Rotation 2 0 100 Yaw 4V 50

0 200 400 600 800 1000 1200 1400 16-0 1800 2000 Sample (128=1 sec)

0 50 0 Roll 0 50 S-100

0 200 400 600 800 1000 1200 1400 !600 1800 2000 Sample (128=1 sec) 4)20 PNch 0

< S)-200- 200 400 600 800 1000 1200 1400 1600 1800 2000 Sample (128=1 secl Intemal: test 2 Elbow Rotation 2 o 100 50 Yaw C 220

i '20

-40 200 400 600 800 1000 1200 1400 1600 1800 2000 Sample (128=1 sec) 00 0 40 520 Roll C 0 00

-80 40 200 400 600 800 1000 1200 1400 1600 1800 2000 Sample (128=1 sec)

040 Pitch 20

-40

60 200 400 600 800 1000 1200 1400 1600 1800 2000 Sample (128=1 sec) T 60 Internal: teat 3 Elbow Rotation 4 40 0-600 Yaw

S20 0 -20 -40 .~0 500 1000 1500 2000 2500 3000 Sample (128=1 sec)

0) 40 T 20 RoM 0) 0-20 Q) -40 S-60 < 0 500 1000 1500 2000 2500 3000 Sample (128=1 sec) 06 40 Pitch 020 v0 W -20 ~40 < 0 00 1000 1500 2000 2500 3000 Sample (128=1 sec)

Figure 3-17: Elbow rotation example from each subject including yaw, roll, and pitch,

The data for each test subject shows the compound motion required to "bend" the ADS arm. Not only does the pilot's arm flex and extend, it also initiates a roll of the forearm. Given the relationship in joint mechanics between the upper and lower

55 M

arm, coupled with the actual placement of the IMU, pitch is less significant than the numerical values produced.

3.6.2 Arm Raise

As with the elbow rotation, the arn raise analysis consists of 60 movements. Figure 3- 18 illustrates a typical graphical output from an arm raise movement group.

Typical Arm Raise 160 shoulder angle V V peak 140 a valley

120 /

100

W, 80

C: 60 Ik

40

20

U 0 / 0 U a M M

-20 L 0 500 1000 1500 2000 2500 3000 3500 Sample (128=1 sec)

Figure 3-18: Typical arm raise angle output of a movement group with five repetitions. The difference between the peaks and valleys represent maximum shoulder movement.

Arm movement is generally smooth, deliberate, and well paced. Pilots do not rush and move the ADS arm as they would under normal working circumstances with a 3-4.5 second period of motion. As seen in Figure 3-18 and consistently among all pilots, each individual movement does not always start with the same arm extension,

56

/ varying by a few degrees each period. Shoulder flexion /extension angles are a result of the magnitude of change between related peaks and valleys and not from a straight maximum determination from the zero x-axis. The small jumps at the peaks and valleys are likely due to a slight shift in arm position within the ADS arm at the beginning and end of the arm raise. This may be coupled with effects from the APDM smoothing algorithm and numerical analysis. Otherwise, this phenomenon is not further investigated due to its low impact on general ROM calculation. Table 3.5 displays shoulder flexion/ extension angle averages over the duration of the experiment. As discussed in Section 3.4, pilots move the joint through maximum rotation. The maximum shoulder flexion/extension angle is 170 degrees from Test Subject 1.

Table 3.5: Arm raise averages over three test subjects.

Mean Angle (deg) Subject 1 145.1 Subject 2 129.2 Subject 3 127.0 Subject 4 -- Total Avg 133.8 Stdev 18.4

Given that all hull-mounted IMUs are on the ADS arm alone, ADS shoulder flexion/ extension data can not be analyzed because there is no relative IMU motion that can adequately compute the arm raise angle. Physical IMU interference concerns impacted hull-mounted IMU placement and it was decided onsite to mount all IMUs in the arm. Future experiments should include more careful consideration of an ADS hull-mounted torso IMU.

3.6.3 Pressure Sensor Correlation

Using pressure pad data processed by Wilkins (2016), kinematic motion from this study is coupled with force data to qualitatively understand how the pilots manipulate the arm. Data from Test Subject 3 best illustrates the motion pattern in both the elbow rotation and arm raise movement groups. The magenta lines on the Figures

57 below indicate a separation of the five individual motions and the green line marks the peak flexion extension angle achieved.

______F Force 100 - -- ______1

60

o 40 L- 20

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

Arm Angle 60 60 O"40-

20 0- -20 -- 6 1- -40 20 21 0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 16 19 Time (s)

Figure 3-19: Test Subject 3 elbow rotation data for a typical movement illustrating two distinct force application periods of flexion (higher force) and extension (lower force).

T Force 80 -

60

40 0- 20-

-12 0- 1 24 25 26 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Time (s)

Arm Angle 150

E? 100-

50

0

-50 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 3-20: Test Subject 3 arm raise data for a typical movement illustrating two distinct force application periods of flexion (higher force) and extension (lower force).

Seen in both Figures 3-19 and 3-20, Test Subject 3 applies a larger force during

58 flexion and lower force during extension. The distinction between motion phases is expected and highlights the unique ADS arm mobility. First, in both cases gravity assists the ADS arm and pilot in extension. As the arm flexes, and then prepares for extension, the pilot releases contact with the ADS hull and contacts the opposite side of the suit to apply smaller contact force to extend or lower the arm. Since both the elbow rotation and arm raise contain elements of downward motion, gravity provides some assistance. Accordingly, if the arms were neutrally buoyant, there would be more equal application of force both up and down. Therefore, these charts indicate the ADS arms are slightly negatively buoyant. Second, there is a longer application of force, or higher impulse, for the arm raise movement as a result of a longer motion (higher flexion/extension range) and movement of a larger (greater frontal area) body through the water. For the elbow rotation, only the lower portion of the arm moves, and for the arm raise, the entire arm raises and lowers through the water column. Overall, the combination of IMU and pressure pad data illustrates a more compre- hensive understanding of motion inside the ADS. A thorough quantitative explanation of forces and detailed exploration of coupled pressure pad and IMU data during elbow rotations and arm raises can be found in Wilkins' 2016 thesis.s

59 THIS PAGE INTENTIONALLY LEFT BLANK

60 Chapter 4

Military Diving Data

4.1 Overview

In an effort to inform the team of researchers and engineers participating in the Navy STTR N13A-T029, a numerical assessment of military diving analyzes past diving history and identifies how a next-generation ADS may fit into diving operations. Diving logs are parsed and charted by depth, type, service, apparatus, and then compared against diving related mishaps.

4.2 Background

In 2008, the US military transitioned to a central, electronic dive logging system called the Dive Jump Reporting System (DJRS). Meant to increase visibility on diving trends and maintain an accessible system to view dive currency, the system is a great asset to the dive community and its leadership at the same time creating a very accessible repository of diving data. In the summer of 2015, the largest data draw in DJRS history was requested via a Freedom of Information Act (FOIA) application: every electronic dive log since DJRS implementation and every dive related mishap. Honoring the FOIA, the Naval Safety Center, owner and maintainer of the system, provides 768,851 dive log entries, 39 mishap reports 2008-2015 (partial year), and 343 mishap reports 1960-2007. In addition to the DJRS data, ADS dive logs from the

61 Portsmouth Naval Shipyard further strengthen the dataset. Saturation Dive logs are not included due to inaccessibility. Six years' worth of military dive history inform the following quantitative assessment of military diving operations. In an effort to share information with the STTR research team and the greater div- ing community, the interactive online plotting tool, Plotly (https://plot.ly), is used to create interactive, editable, and shareable charts. All charts are copied in static form to this study, but also made available online at http://www.militarydivingdata.com

(or http://divingresearch.scripts.mit.edu/militarydivingdata after December 2016) for an interactive experience.

4.3 Quantitative Methodology

4.3.1 DJRS Dive logs

As described in section 4.2, DJRS is an electronic record keeping online software tool for military diving and skydiving operations. Since 2008, DJRS requires that dive commands shift from the old method of dive log record keeping. All commands have since adopted the process, which provides a regulated, controlled, and accessible repository of all military diving logs. Although the method in which dive logs are recorded in the online software is regulated with drop down menus and text con- trolled input, it is not without some user input error as discussed in more detail in section 4.4.1. System administrators at the Naval Safety Center provide the diving data file in Microsoft Excel® format which contains 768,852 rows and the following column headings:

" Dive Date

" Calendar Year (CY)

" Reporting Service

" UIC (command identifying code)

" Command Name

62 " Repet Group (letter representation of residual Nitrogen in a diver post dive in accordance with US Navy (2008))

* Left Surface (date/time)

" Left Bottom (date/time)

* Reached Surface (date/time)

" Dive Purpose Description

" Apparatus Description

* Water Temperature

" Schedule Time

" Schedule Depth

* Maximum Depth

" Surface interval

" Description and applying notes

Given the detailed nature of the included data, individual dives and associated repetitive dives could be recreated and understood. This raw dataset provides an op- portunity to compare many aspects of diving operations based on depth bands (e.g., Dive Purpose, Apparatus, Service, and overall number of dives). Due to the size of the dataset, manipulation in Excel is too cumbersome. Therefore, the diving dataset is imported to MATLAB for faster computing analysis and Plotly for visual display. In MATLAB, using "category" functionality, data is easily parsed and prepared for graphing based on user input. Using the Plotly Application Programming Interface (API), prepared datasets are uploaded to the Plotly web interface and visually ad- justed for maximum clarity.

63 4.3.2 Diving related Mishap Reports

Mishap data dates as far back as 1960, totaling 382 diving related mishaps. The Excel file contains the following column headings:

* Dive Date

" Dive Serial Number

" Left Surface (date/time)

* Reached Surface (date/time)

" Apparatus Description

* Decompression Type

" Dive Platform

" Water Temperature

" Final Stage Depth

" Schedule Depth

" Diving Accident Type

" Mishap narrative

Mishap data is analyzed similar to the methodology described in section 4.3.1 and is used to correlate mishap data and number of dives to find a mishap rate, as well as identify any trends related to apparatus use and diving related illnesses. The mishap narrative is an invaluable dataset feature when assessing outliers and inconsistencies.

4.3.3 ADS Diving Logs

ADS Diving Logs provided by Portsmouth Naval Shipyard are not maintained in DJRS. Although the ADS is operated by Navy divers, much of the input to DJRS is not applicable and therefore, the ADS is treated more like a "submersible" than

64 a "human diver." Provided data on the Navy's four ADS 2000s is formatted and added to the overall dataset to supplement the DJRS dive logs to provide deep-diving recognition and system utilization.

4.4 All Military Diving Breakdown

Figure 4-1 is a traditional histogram showing all military dives since 2008. All services, types of diving rigs, and tasks are represented. Some technical notes are (1) dives are broken into incremented 10 ft "bins," starting from the left limit; (2) the blue columns represent the total number of dives on a logarithmic scale; and (3) the orange line represents the Empirical Cumulative Distribution Function (ECDF) of dives achieved to a given depth. The selection of 10 ft depth bins is meant to closely relate the histogram to typical depth intervals on Navy Diving Tables (US Navy, 2008). Note the depth axis stops at 310 ft, above which, the ADS is the only rig used, absent of saturation diving (data not included).

65 Military Wide Diving Distribution 2008-2015

Note: Depth bins ae 10 in rements -tarnig a' ne lef* hm t s- 100

100k 4 90% of all dives are 60 ft or less total Mill tary Dives -90% 2- ECDF

80%

10k-s- *Iq

-60% 5earky 600% of ali Iiilr~e- ierm (flr )-i,~hv

1000 -504 a) -0 E) -D -40",'

100- z eal faldvshv Ulintd.0 ietm <0t -30%

10- -10%

0 10 20 30 40 50'I C0 90 ljy 110 120 130 140 150 160 170 10 190l0t 21 220 230 24) 250 20 270 2180 290 300 310 Depth (It'

Figure 4-1: Histogram of all military diving distribution 2008-2015 with ECDF overlay. The dataset indicates an average dive depth of 27.8 ft. This number is driven by high instances of 10-29 ft dives (20 ft dives representing the dataset median). Addi- tionally, nearly 60% of all dives have "unlimited" dive time and 90% of all completed dives are less than 60 ft. "Unlimited" dive time refers to an unrestricted time al- lowance at depth based on the body's ability to rid itself of inert gasses without an incurred decompression stop US Navy (2008). With 90% of military diving occurring less than 60 ft, this shows the shallow water diving tendency of the military. Figures 4-2 and 4-3 illustrate "how" the military dives. DJRS allows the documen- tation of every dive by a defined list of "Dive Descriptions." Each bin is normalized to the total number of dives within the given depth bin as shown in Equation 4.1.

Z Dives of a specific Description (4.1) E All Dives of all Descriptions Figure 4-2 illustrates a group normalized percentage for each given dive type. For example, between 260 and 269 ft, 50% of time is spent on Experimental Testing (dark green), and the other 50% is spent on Search (royal blue). Most depth bins are dominated by training. Training (Diver), Supervisor Training, and Student Training represent 57% of all diving depths less than 190 ft. Some points of interest are (1) the military diving community spends a considerable amount of time on Experimen- tation, specifically at deeper depths; (2) Pressure Testing for dive candidates and Recompression treatments make an appreciable appearance in the 60 ft depth bin; and (3) Requalification dives are generally very shallow (10-19 ft).

67 Military Wide Diving Distribution in Color

lon Work-up Dive U/W Construction Ops Training (Diver) . Supervisor Training Student Training 80%- Ships Husbandry/Repair Security Swim Search Salvage/Recovery

- iRoutine Working Dive 60%- Research Requalification Recomp Trtmt Pressure Test Other Instructor Safety Obsrvr 40%- m i nInspection/Survey Indoctrination Ice Diving INSD Tndr/Sur-D Dives Humntr Trtmt 20%- -- Force Protection

Experimental N MW EOD Opt

_W Combat Swimmer Ops/Training 1 Clinical Hyperbaric Trtmt 0% .- 200 210 220 230 240 250 260 270 280 290 300 310 6 Aids to Navigation 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190

Depth (ft)

Figure 4-2: Group normalized histogram of all military diving color coded by job description. Figure 4-3 shows all diving work once all training and non-operational diving types are removed. The shallower depths are dominated by harbor work: U/W Construc- tion, Ship's Husbandry, Security Swim, Routine Working Dives, Inspection/Survey, and Force Protection. As depth increases, the dive profile is dominated by Salvage, Search, and Explosive Ordnance Disposal (EOD) work. This chart is best viewed on- line (www.militarydivingdata.com), allowing chart customization by toggling legend descriptions on and off.

69 Military Wide Diving Distribution in Color

100%-

U/W Construction Ops

80%- UShips Husbandry/Repair Security Swim Search Ir-- Salvage/Recovery _0 Routine Working Dive

0 7- T 0 40%- 0

U 40%/o Inspection/Survey

)!xorl

20%1 I Force Protection

EOD Ops

0% r Clinical Hyperbaric Trtmt 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 1 Aids to Navigation

Depth (ft)

Figure 4-3: Group normalized histogram of all military diving color coded by job description. Training specific diving types are toggled off (grayed out) to highlight the changes in normalized distribution of non-training diving types. ~-~-- -~

Figure 4-4 breaks out the data by Service affiliation. As expected, the Navy (blue) dominates the number of dives completed at each depth. Army Divers (green) make up the next highest percentage of dives at almost all depths. This chart represents all types of diving Descriptions on all types of diving apparatus.

Military Wide Diving Distribution By Service

0 Navy E Marines

-- Coast Guard O Army Air Force S IUNK

0 0 20 30 so v 70 8 100c0 10s 1 30 40 05 l 17 0 18r 10 20 2 0( 260 270 2 80 290 700 310

Deptn (ft) Figure 4-4: Group normalized histogram of all military diving color coded by service affiliation.

Figures 4-5 and 4-6, segment the diving data by apparatus type. This is the first time the ADS dive history is visible, although at a very small fraction. The pie chart in Figure 4-5 represents the average apparatus use per year. The group normalized chart in Figure 4-6 shows how apparatus use relates to depth. The "Oxygen Re- breathers" category represents MK-25 variants, "Mixed Gas Rebreather" category represents MK-16 (Air and Helium diluents) or similar variants, "Other" category includes experimental or similar rigs, "Surface Supplied" category represents MK-21 and KM-37 similar variants (mixed gas and air), and SCUBA and Chamber categories represent their namesake. The combined categories provide a more informative chart without excessive granulation. Actual dive logs, for example, may include different varieties of MK-16 diving apparatus, but the intent of this chart is to see general apparatus breakdown.

71 I N06-

Military Diving By Apparatus

ADS 0.0497%

P" I Other 0.236%

Surface Supplied 19.9%

Mixed Gas Rebreather 3.91%

Figure 4-5: Pie chart of all military diving segmented by diving apparatus type.

Military Wide Diving by Apparatus

E Surface Supplied 0 -IPSCUBA N Oxygen Rebreather W Other R Mixed Gas Rebreather 0 Chamber R ADS

10 20 '?0 10 6 70 80 90 100 11D 130-0 14 150 ISO 1IN 180 110 200 210 220 230 240 250 26i 270 280 20 300 316

Depth (1t)

Figure 4-6: Group normalized histogram of all military diving color coded by diving apparatus type.

As seen in Figure 4-6, SCUBA (orange) dominates shallower depths and trails

72 off until its upper authorized-use limit: 190 ft (US Navy, 2008). The five SCUBA dives at 200 ft recorded by the Navy Experimental Dive Unit (NEDU) are assumed to be related to experimentation efforts. Mixed Gas Rebreathers and Surface Supplied diving becomes dominant as depth increases demonstrating the necessity for exotic gas diving. Few dives (536 total, less than 100 per year) are completed at depths greater than 200 ft; therefore, the apparatus use percentages are more variable and sensitive in this region.

Figure 4-7 segments the dives by year. Given that 2008 represents the start of DJRS, and only partial data exists for 2015, the following chart represents the dive breakdown from 2009 through 2014. Although the number of dives increases steadily over the years, all previous analyses are annually consistent. On average, as a combined military diving community, more than 115,000 dives are logged per year. Over 45 dives occur every working hour (assuming 250 day work year and 10 hour

work day).

Military Wide Diving per year 2009-2014

20k- Average Number of Dives per Year 115,241

100k

80k

E 60k

,Tk

010 201 1 212013 2014 Year

Figure 4-7: Histogram of all military diving segmented by year.

73 4.4.1 Unacceptable Dive Logs

There are inconsistencies with a small portion of dive logs, identified during pass analysis of the raw diving dataset. Specifically, SCUBA (lives are frequently displayed above 200 ft. Upon further inspection, there are nearly 13,000 misreported dive records ( 1.5% of all logs). Those misreported dive records either report a Max Depth greater than Schedule Depth or do not record a Schedule Depth. Schedule Depth refers to the depth used to calculate how long a diver must remain topside to allow the gasses in their body to normalize and should be greater than or equal to the Max Depth. In some instances, it is easy to identify a mistype (Schedule Depth of 30 ft, Max Depth of 300 ft). Logs with recorded Schedule Depth of zero can not be verified against max depth, and therefore indicate an untrustworthy dive log. All questionable dive logs are excluded from this analysis.

4.5 Diving Related Mishaps

The following Navy and Marine Corps diving mishap data gives the military diving community enhanced situational awareness. Most years contain less than 10 dive mishaps with an annual mishap rate (mishaps per number of dives in a given depth band) of less than a hundredth of a percent. While experiencing a limited number of diving related mishaps is favorable, this absence of data makes forming solid statistical analysis less impactful. Still, a low mishap rate speaks highly of the level of training, supervision, equipment, and technical prowess within the Navy and Marine Corps diving community. Mishaps reported to the Naval Safety Center represent an "injury, recompression therapy, or death resulting from an incident occurring while breathing compressed gases .... before, during, or after entering or leaving the water" (OPNAVINST 5102.1). The Naval Safety Center report contains 39 mishap reports (2008-2015). Two known mishaps that received media attention from February 2013 (not provided by the Safety Center) are added to the dataset to ensure its completeness. One additional report is corrected based on an erroneously reported max depth corroborated in Dive

74 Narrative. Overall, 41 mishaps are analyzed. Figure 4-8 shows the mishap histogram (2008-mid 2015) with associated diving mishap type (illness) illustrated below. A strong statistical correlation, R, (+0.85) exists between number of dives and number of mishaps. Meaning, there exists a posi- tive and predictive linear relationship between the number of dives and the number of mishaps. Related, there is a moderate negative correlation (-0.65) between mishaps and depth, meaning more mishaps at shallower depths. As the bottom chart of Figure 4-8 shows, the Navy and Marine Corps experience a variety of diving related illnesses. Well known diving illnesses such as Decompression Sickness (DCS) and Pulmonary Over Inflation Syndrome (POIS), specifically Arterial Gas Embolism (AGE), are found experientially through many depth bands illustrating the ever present danger of illnesses unique to diving. Illnesses such as DCS and AGE require immediate treatment via recompression chamber. A complete description of diving illnesses and required treatment can be found in the Navy Dive Manual.

75 Diving Related Mishaps by Casualty Type since 2008

8-

-c

0- 5- V) 4-

3- Totals BAROTRAUMA z CHEM/TOXIC EXPOSURE 1 1 DCS I DROWNING 0 10 20 30 .0 50 60 70 80 9f 100 0 I0 I 150 160 170 HYPERCAPNIA MISSED DECOMPRESSION 100%- EAR DROWNING --A 0- 80% * ETINAL EDEMA A TERIAL GAS EMPOLISM 60% * DCS I1 0 a 40%

20%

0 10 20 30 4 50 60 70 80 90 110 120 130 140

Depth (ft)

Figure 4-8: Histogram of Navy and Marine Corps diving mishaps 2008-mid 2015 with related diving casualty type. Figure 4-9 represents the Navy and Marine Corps diving histogram (logarithmic scale) overlaid with the Mishap Rate during the last six years. Note the Mishap Rate can be identified by the right y-axis and only reaches a maximum of 0.53% in the 150-159 ft depth bin. Although the Mishap Rate value is very small, there does exist a statistically significant (p < 0.05) change in Mishap Rates from the 0-99 to 100-170 ft depth. This statistical significance requires further inspection. The majority of the mishaps in the 100-170 ft depth band (9 of 14) are air diluent MK-16 dives, including one fatality. 3 of 14 dives are HeO 2 diluent MK-16 dives. The last mishaps in this depth band are two fatal SCUBA dives in February of 2013. Overall, nearly 80% of mishaps in this depth band occur while breathing air.

77 Navy and Marine Corps Diving Distribution with Mishap Rate 2008-2015

3.00,

100k- U Total Navy and Marine Corps Dives

5- Mishap Rate -2.50%

10k-

a) 2.00% (U

no 0 1000-

a) 5- 1.50% C 0 100 a) ~0 E 3 .00% z I 10-

0.50%

1- a A 'I 22n 230 '40 -,n (;0 27,' 28O '90 300 1 ) 10 20 3( 40 5) 60 ) 0 80 9) 100 '1( 120 'Y 140 150 1606 7f 01020o -in

Depth (ft)

Figure 4-9: Navy and Marine Corps diving distribution with Mishap Rate 2008-2015. Figure 4-10 shows further analysis of the association between apparatus and mishaps. The data includes 22 Mixed Gas Rebreather mishaps across the depth band, 14 02 Rebreather mishaps between 10 and 30 ft, 4 SCUBA mishaps, and 1 Surface Supplied mishap. The supplemental notes on the chart indicate the maxi- mum individual Apparatus Mishap Rate. Generally very low, the Apparatus Mishap Rate provides perspective for the "spikes" seen in Figure 4-10. The 50% SCUBA Mishap Rate in the 150-159 ft depth bin is an outlier, but highlights the rarity in which SCUBA is used for deep diving. Only 60 logged SCUBA dives since 2008 occur at 150 ft or greater, with 2 reported mishaps (overall 3.3% Mishap Rate for "deep" SCUBA dives).

79 Mishaps by Apparatus

7- of 02 Rebreather Dives in this depth band resulted in mishaps 0.01% * Chamber * Mixed Gas Rebreather * Other * Oxygen Rebreather 1.4% of MG Rebreather dives in this depth band resulted in mishaps SCUBA 5- 1I * Surface Supplied 4- 0

EJ 50% of SCUBA dives in this depth band resulted in mishaps 3' I1 zo 2-

I-

I I I I- 1 rn) M I M Is M 111 120I '50 160 170 80 90 100 I 20 30 40 50 60 70 I luo 10 I I 0 1 I Depth (ft)

Figure 4-10: Navy and Marine Corps diving mishaps by apparatus type. Chapter 5

Conclusions

Overall, this study was able to achieve the following goals:

e Demonstrate the effectiveness of limb kinematic assessments using IMUs.

e Establish ADS pilot ROM for elbow and shoulder flexion/extension movements.

e Provide a framework for follow-on new joint assessment using IMUs.

e Strengthen project team knowledge of military diving operations with compre- hensive quantitative assessment of diving operations.

Much like the spacesuit studies using IMU kinematic assessments referenced in Chapter 1, this study proves a similar effectiveness in determining pilot mobility. Specifically, the determination of simple elbow and shoulder flexion/extension range of motion baselines typical pilot arm mobility. Future joint systems should aim to exceed these values in order to deliver greater pilot ROM and allow more natural work. With an elbow flexion/extension near 80 degrees and shoulder flexion/extension near 134 degrees, the ADS restricts pilot mobility over unencumbered human mobility. Furthermore, arm flexion/extension in the ADS is a nonplanar motion, requiring elements of rotation to achieve a change in arm angle. Extra motion of any kind contributes to fatigue and, in the case of the ADS, requires a high level of training to learn effective and efficient movement. Designing a suit that more closely mimics natural human motion could increase the suit's capability and minimize training.

81 This study's testing framework effectively demonstrates the concept of kinematic evaluation. Improvements to the test routine should include: the use of more test subjects, more defined motion patterns, and real time data view using custom or third- party software scripts. In order to gain further clarity and confidence in the numerical angle assessment, future experiments should include more in-depth alignment and dy- namic calibration procedures similar to those applied during medical-grade precision limb-testing. The inclusion of custom or third-party real time calculation software can also provide valuable movement visualization and problem identification. Addi- tionally, redistribution of ADS hull mounted IMUs should be considered to more fully monitor ADS arm motion. Placement of an IMU on the the ADS torso would allow ADS shoulder joint assessment during the arm raise movement. Beyond the arm, the hip and knee movement of the pilot and ADS should also be included in follow-on studies.

Data analysis of military diving reveals a relatively shallow diving military. The majority of dives made by the military are less than 60 ft in depth. The greatest number of diving related mishaps also occur in that range for the Navy and Marine Corps. Diving related mishaps are rare, but have the potential to cause lasting, life-threatening effects with the continued reality of diving illnesses such as DCS, AGE, and drowning. The ADS, on the other hand, has not been involved in any diving related mishaps nor does it expose the pilot to the threat of diving illnesses. Independent of the ADS, the military does dive at the upper allowable limits (200- 300 ft), but completes less than 100 dives per year at this depth. The incredible pressures and variable physiological effects at deep depth tend to restrict deep diving operations. With the development of a more natural moving, lightweight, shallow (<1000 ft) diving ADS there may be an increase in deep diving without the threat of diving illness.

82 5.1 Future Work

Exposure to the world of IMU kinematic evaluation and its effectiveness has prompted the consideration and thought of future applications.

First, continue to explore IMU and other sensor based kinematic assessments. The methodology applied in the study allows a relatively low cost evaluation of the ADS pilot and the suit. Use this concept to evaluate other aspects of the current ADS as well as next generation joint designs.

Second, consider fully developed third-party systems with real time display. Many software suites on the market include full hardware and software systems designed for kinematic testing. Many of the product solutions can be scaled to a project's needs. Computation and analysis time can be dramatically reduced with more focus placed on effective application and results analysis. With little training, a system can be employed on a subject and full body assessment completed, all in real time.

Lastly, beyond mobility assessment, consider the use of an IMU based sensing system as a means of communication and situational awareness for supervisors. Top- side ADS diving supervisors rely on voice and video to understand the condition of their pilot and progress of a job. With the use of an integrated, full body IMU system employed on the ADS pilot and/or suit, the supervisor can directly see limb motion, orientation, and even navigational position. With limb motion viewing, the supervisor can assess job progress without the interference of voice communication or identify troubling limb positioning. Furthermore, this capability can assist supervi- sors in training new pilots, quickly identifying inefficient movements. With orientation awareness, the supervisor can assist the pilot with project problem solving, attitude correction, and identification of potential problems. Finally, although not discussed in this study, IMUs can be used to determine relative navigational position. Often used in underwater and airborne vehicles, a hull mounted IMU can give the supervi- sor navigational position information on the ADS in the water column. This would allow the supervisor to assist the ADS to move to a project location faster and more accurately, especially in poor visibility conditions. Overall, a suite of IMUs can bring

83 a 1new( dimension in situational awareness for ADS diving supervisors.

Figure 5-1 gives a conceptual view of the integration of an IMU monitoring system.

Illustrated on the center console, the IMU real-time bodI position display (an be easily integrated with an existing third party system and utilization of existing data connections available in the tether.

Figure 5-1: Concept of full IMU body awareness as a means of conunnication for ADS diving supervisors shipboard. Note the skeleton IMU rendering on the center console representing pilot orientation and limb position.

84 Appendix A

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 + 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 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 1X). - 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). - 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 on the pilot and of the ADS would help elucidate these differences. As stated previously, this be experiment uses similar methods to those used in Space Suit evaluation and were found to effective in identifying human-suit interaction. This data will be used not only to determine to biomechanical differences, but also to help find points of maximum and minimum movement analyze the pressure profiles. Furthermore, IMU data will be matched with video data to improve of the and explain results visually. Sensors will be placed on the RIGHT lower and upper arm 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. with middle and wrist) on the RIGHT arm, positioned insider "spacer areas" and not to interfere pilot movement or pressure mat contact.

- Commercially available APDM Opal inertial measurement unit (IMU) * 3 pilot mounted sensors : upper arm, lower arm, and chest e 3 internally hull-mounted sensors: upper and lower arm * Each is 4.8x3.6x1.3 cm (lxwxh) and weighs less than 22g * Lithium Ion battery at 3.7V nominal 8h the current is 56 mA. With e Electrical capacity is 450mAh. Assuming that it can last minimum 16 hours of operation, the current is 28mA include aluminum base to e The worst case scenario is venting of the battery. Safety precautions 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 will be This test will be performed with as many subjects as time permits. The following tasks for performed while suited in the ADS. Each task will be repeated through 5 repetitions in 4 groups minutes and a total of 20 repetitions. After each movement is performed, the subject will rest 2 and rest subjective data will be collected. The subject will then repeat the movement sequence period four additional times.

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

Elbow Rotation

87 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 THIS PAGE INTENTIONALLY LEFT BLANK

92 Bibliography

Wei Sin Ang, I-Ming Chen, and Qilong Yuan. Ambulatory measurement of elbow kinematics using inertial measurement units. In 2013 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM), pages 756-761, Wollon- gong, Australia, 2013.

APDM. APDM - About APDM, 2016a. URL http://www.apdm.com/about-apdm/.

APDM. APDM - Wearable Sensors, 2016b. URL http://www.apdm.com/ wearable- sensors/.

Leslie G. Baehr. The Exosuit Comes Aboard: Cutting-edge diving apparatus gets first archaeological mission, 2014. URL http://www.whoi.edu/oceanus/feature/exosuit.

Pierre J. Bertrand. Enhancing Astronaut Mobility Spacesuit Kinematics and Interac- tive Space Outreach. Master's thesis, Massachusetts Institute of Technology, 2016.

CODE 205 Deep Submergence Systems Program Office. Scope of Certification Note- book Volume 1 for the Atmospheric Diving System 2000 (ADS 2000), 2005.

Mahmoud Ahmed El-Gohary. Joint Angle Tracking with Inertial Sensors. PhD thesis, Portland State University, 2013.

J. Favre, B. M. Jolles, R. Aissaoui, and K. Aminian. Ambulatory measurement of 3D knee joint angle. Journal of Biomechanics, 41(5):1029-1035, 2008.

J. Favre, R. Aissaoui, B. M. Jolles, J. A. de Guise, and K. Aminian. Functional calibration procedure for 3D knee joint angle description using inertial sensors. Journal of Biomechanics, 42(14):2330-2335, 2009.

Craig Freudenrich Ph.D. How Space Suits Work, 2000. URL http://science. howstuffworks.com/space-suit4.htm.

Douglas E Kesling. Atmospheric Diving Suits - New Technology May Provide ADS Systems that are Practical and Cost-Effective Tools for Conducting Safe Scien- tific Diving, Exploration, and Undersea Research. In Proceedings of the American Academy of Underwater Sciences 30th Symposium, pages 74-76, Dauphin Island, Alabama, 2013.

93 Ryan L Kobrick, Christopher E Carr, Forrest Meyen, Ana R Domingues, Prof Dava J Newman, and Shane E Jacobs. Using Inertial Measurement Units for Measuring Spacesuit Mobility and Work Envelope Capacity for Intravehicular and Extrave- hicular Activities. In 63rd InternationalAstronautical Congress, pages 1-9, Naples, Italy, 2012.

Emory Kristof and Joseph B. MacInnis. Exploring a 140-year-old Ship Under Arctic Ice. National Geographic Magazine, page 104, 1983.

Irvin Hussein Lopez-Nava and Angelica Munoz-Melendez. Estimation of Angles from Upper and Lower Limbs for Recognition of Human Gestures using Wireless Inertial Sensors. Technical report, Instituto Nacional de Astrofisica, Optica y Electronica Tonantzintla, Puebla, Mexico, 2015.

MD. Reporter's Notebook: Want to swim with the fishes? Try this swimmable submarine. Machine Design, pages 24-26, feb 2009.

Phil Nuytten. EXOSUIT ADS Factsheet, 2016. URL http://nuytco.com/wp-content/ uploads/2013/06/Exosuit-Specifications.pdf.

Kai Daniel Oberldnder. Inertial Measurement Unit ( IMU ) Technology, Inverse Kinematics : Joint Considerations and the Maths for Deriving Anatomical Angles, 2015.

OceanWorks. HARDSUIT Factsheet, 2016. URL http://www.oceanworks.com/ admin/sitefile/ 1 /files/OceanWorks ADS(5).pdf.

A Reinhardt. Results and Applications of a Space Suit Range-of-Motion Study., 1989.

Tyler Rudick. You down with OTC? It's safety, Hooters Girls (with coupons) & iPad 3 races at mega oil conference, 2012. URL http://goo.gl/fslRjT.

Gregory G Slabaugh. Computing Euler Angles From a Rotation Matrix, 1999. URL http://gregslabaugh.name/publications/euler.pdf.

Mike Thornton, Robert Randall, and Kurt Albaugh. Then and Now: Atmo- spheric Diving Suits. Underwater Magazine, pages 1-8, 2001. URL http://www. underwater.com/archives/arch/maraprO1.01.shtml.

US Navy. US Navy Diving Manual Revision 6, volume 1-5. Naval Sea Systems Command, 2008.

Bong Wie. Space Vehicle Dynamics and Control. AIAA education series. American Institute of Aeronautics and Astronautics, 2 edition, 2008.

Christopher Michael Wilkins. An Experimental Study of the Human Interface with One Atmosphere Diving Suit Appendages. Master's thesis, Massachusetts Institute of Technology, 2016.

94