Spacesuit and Portable Life Support System Center of Gravity Influence on Astronaut Kinematics, Exertion and Efficiency

Home , Mark III (space suit), Yastreb

A thesis submitted to the Graduate School of the University of Cincinnati

in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

In the Department of Aerospace Engineering and Engineering Mechanics of the College of Engineering and Applied Science

November 2015

SIDDHARTH SRIDHAR

B.Tech Mechatronics SRM University, May 2013

Committee Chair: Grant Schaffner, Ph.D.

Spacesuit and Portable Life Support System Center of Gravity Influence on Astronaut Kinematics, Exertion and Efficiency

Abstract

NASA has initiated a series of tests aimed at understanding human physiological and biomechanical effects of spacesuits under a variety of conditions. Though these tests include metabolic rates, ground reaction forces, biomechanics, subjective workload and controllability feedback, the influences (kinematics, exertion and efficiency) of a combined spacesuit and portable life support system (PLSS) center of gravity (CG) during an astronaut’s extravehicular task performance has not been completely understood. The work described in this thesis was aimed at developing a quantitative means of evaluating the influence of space suit and PLSS CG location on astronaut EVA task performance in terms of kinematics (joint angular ranges), exertion (joint torques and muscle forces), and efficiency (comparative work performed). Four CG locations, representing approximate CG extremes for the NASA MK III and Z1 space suits, were evaluated using a combined experimental and computational approach. Three common EVA tasks were studied: object translation, climbing and walking. It was found that the Low-Aft CG was the best for object translation, the High-Forward CG for walking and the Low-Forward CG for climbing.

iii

Acknowledgements

I would like to thank Dr. Grant Schaffner for his guidance and patience in helping me throughout my Master’s research at the Human Systems and Simulation Laboratory, UC. I would like to thank

Shane M. McFarland at the NASA Johnson Space Center for his constant support throughout this project. Many thanks to Protostar Engineering Inc., Cincinnati for supporting my project partially and helping me design and build the CG variation fixture. My heartfelt appreciation to Dr. Kelly

Cohen and Dr. Kristin Yvonne Rozier for serving on my thesis committee. I would like to acknowledge Eric Stetz, my lab mate for all the NASA spacesuit calculations. I owe you one!

Thanks to the Department of Aerospace Engineering and the Department of Physics for providing me teaching assistantship that helped cover my tuition and living expenses partially, Gaurav

Mukherjee for teaching me on how to use the motion capture system, OpenSim and taking me out for beer many times and encouraging me with all his positive energy. Cheers to my HSSL lab mates Prashanth, Satya, Brandon and Anang for bearing with my sense of humor and working together on Open Sim. Special thanks to Rob Ogden and Curtis Fox of Aerospace Engineering for helping us fix anything that broke.

I would like to extend my thanks to Dr. Kristin Yvonne Rozier for offering me a PhD position in her lab at UC and sending me to the fifth summer school on formal techniques, CA. Much appreciation to the organizations in UC in which I was/am a part of – the Indian Students

Association, the Aerospace Graduate Students Association, the Administrative Review

Committee, Accelerating Racial Justice.

Most importantly, I would like to thank my parents, my sister and all my family in India and abroad for their love, motivation and support. Also, I would like to thank my extended family and buddies at Cincinnati: Anusha, Anudeep, Suprabh, Abinaya, Kshitij, Navneet, Santosh, Deepthi, Devesh,

Pallavi, Selva, Anoop, Sarthak, Shishir, Rohit, Rohan, Vishakh, Abhay, Rajit, Robins, Vamshi,

Bhargav, Ranjan, Rohit Dureja, Madhumitha, Supriya, Hari, Bala and Vishwa to name a few, without whom, my grad life wouldn’t have been the same.

Lastly, thanks Starbucks for all the coffee that kept me awake and Zipcar that took me places.

Table of Contents

Abstract ...... ii Acknowledgements ...... iv Table of Contents ...... 1 List of Figures ...... 2 List of Tables ...... 4 List of Acronyms ...... 5 1 Introduction ...... 6 1.1 Background and Motivation ...... 6 1.2 Research Goals and Approach ...... 7 1.3 Thesis Outline ...... 8 2 Literature Review...... 9 2.1 Spacesuits ...... 9 2.2 Related NASA Studies ...... 12 3 Experimental Design ...... 16 3.1 Center of Gravity Variation Fixture ...... 16 3.2 Methods ...... 20 3.2.1 Subject Preparation ...... 21 3.2.2 Experimental Protocol ...... 25 3.3 Post Experimental Analysis ...... 32 3.3.1 Tracking the Motion Capture Data ...... 32 3.3.2 OpenSim Analysis ...... 33 4 Results ...... 37 4.1 Differences in kinematics ...... 37 4.2 Exertion ...... 43 4.2.1 Joint Torques ...... 43 4.2.2 Joint Power...... 49 4.2.3 Muscle Force ...... 52 4.3 Efficiency of task performance ...... 57 4.4 Combined Results ...... 60 4.5 Rate of Perceived Exertion ...... 61 5 Discussions and Conclusions ...... 63 6 Future Work ...... 65 References ...... 66 Appendix A ...... 68

List of Figures

Figure 2.1: Extravehicular Mobility Unit………………………….……………………………..12 Fig. 3.1: A subject wearing the CG variation fixture……………………….……………………19 Fig. 3.2: CG variation fixture……………………………………………….……………………19 Fig 3.3: Comparison of Target CG and fixture + subject CG……………….…………………...19 Fig 3.4: Four CG locations of interest……………………………………….…………………...20 Fig. 3.5: Frontal and left side view of the model with the markers……….…………………….....22 Fig. 3.6: Rectus femoris sensor placement………………………………………………………..24 Fig. 3.7: Vastus lateralis sensor placement……………………………………………………….24 Fig. 3.8: Tibialis anterior sensor placement……………………………………………………....24 Fig. 3.9: Gastrocnemius medialis sensor placement……………………………………………...24 Fig. 3.10: Biceps femoris captus long sensor placement……………………………………….....25 Fig. 3.11: Erector spinae ileocostalis sensor placement……………………………………….….25 Fig. 3.12: Motion Capture Volume……………………………………………………………….26 Fig. 3.13: MVC for the rectus femoris and the vastus lateralis muscles……………………….....27 Fig. 3.14: MVC for the gastrocnemius muscle…………………………………………………...27 Fig. 3.15: MVC for the biceps femoris muscle…………………………………………………...28 Fig. 3.16: MVC for the erector spinae ileocostalis muscle………………………………………..28 Fig. 3.17: Static Pose……………………………………………………………………………..29 Fig. 3.18: Object Translation (Low-Aft CG) as performed by a subject………………………….30 Fig. 3.19: Climbing (High-Front CG) as performed by a subject…………………………………31 Fig. 3.20: Walking (High-Aft CG) as performed by a subject…………………………………….32 Fig. 3.21: Schematic of the Computed Muscle Control Algorithm……………………………….35 Fig. 4.1 (a): Hip Flexion Joint Angle Comparison with STS Spacesuit Joint Mobility Range……39 Fig. 4.1 (b): Average Minimum Hip Flexion Joint Angle………………………………………...40 Fig. 4.1 (c): Average Maximum Hip Flexion Joint Angle………………………………………...40 Fig. 4.2 (a): Knee Joint Angle Comparison with STS Spacesuit Joint Mobility Range…………...41 Fig. 4.2 (b): Average Minimum Knee Joint Angle………………………………………………..42 Fig. 4.2 (c): Average Maximum Knee Joint Angle……………………………………………….42 Fig. 4.3: Hip Flexion Joint Torque – Climbing…………………………………………………...44 Fig. 4.4: Hip Flexion Joint Torque – Object Translation……………………………………….…44 Fig. 4.5: Hip Flexion Joint Torque – Walking……………………………………………………45 Fig. 4.6: Knee Joint Torque – Climbing…………………………………………………………..45 Fig. 4.7: Knee Joint Torque – Object Translation………………………………………………...46 Fig. 4.8: Knee Joint Torque – Walking…………………………………………………………...46 Fig. 4.9: Average Max Joint Torque – Hip Flexion……………………………………………….47 Fig. 4.10: Average Max Joint Torque – Knee…………………………………………………….48 Fig. 4.11: Maximum Joint Torque Results………………………………………………………..49 Fig. 4.12 (a): Average Max Joint Power – Hip Flexion………………………………………….50 Fig. 4.12 (b): Average Max Joint Power – Knee………………………………………………….51 Fig. 4.13: Maximum Joint Power Results ………………………………………………………..52 Fig. 4.14: Average Maximum Muscle Force – Rectus Femoris…………………………………..53 Fig. 4.15: Average Maximum Muscle Force – Vastus Lateralis………………………………….54 Fig. 4.16: Average Maximum Muscle Force – Biceps Femoris Captus Long…………………….54

Fig. 4.17: Average Maximum Muscle Force – Tibialis Anterior…………………………………55 Fig. 4.18: Average Maximum Muscle Force – Gastrocnemius Medialis…………………………56 Fig. 4.19: Average Maximum Muscle Force – Erector Spinae Ileocostalis………………………56 Fig. 4.20: Maximum Muscle Force Results………………………………………………………57 Fig. 4.21: Average Work – Hip Flexion…………………………………………………………..58 Fig. 4.22: Average Work – Knee…………………………………………………………………59 Fig. 4.23: Joint Work Results…………………………………………………………………….60 Fig. 4.24: Combined Results……………………………………………………………………..61 Fig. 4.25: Average Rate of Perceived Exertion…………………………………………………...62 Fig A1: Corlett and Bishop Discomfort Scale…………………………………………………….70

List of Tables

Table 3.1: NASA Spacesuit CG declassified data………………………………………………..16 Table 3.2: Combined spacesuit and subject CG…………………………………………………..17 Table 3.3: Combined spacesuit and subject CG…………………………………………………..19 Table 3.4: De-identified anthropometric measures of the subjects……………………………….21 Table 4.1: Hip Flexion Joint Angle Comparison with STS Space Suit Mobility Range…………..37 Table 4.2: Knee Joint Angle Comparison with STS Space Suit Mobility Range…………………38 Table A1: Rate of Perceived Exertion Scale……………………………………………………...68 Table A2: RPE Data Log…………………………………………………………………………69 Table A3: Rate of Perceived Exertion Experimental Data………………………………………..69 Table A4: Corlett and Bishop Discomfort Scale Data Log……………………………………….71 Table A5: Corlett and Bishop’s Discomfort Scale Experimental Data…………………………...71

List of Acronyms

PLSS………………. Primary/Portable Life Support System CG………………..... Center of Gravity EVA……………….. Extravehicular activity NASA……………… National Aeronautics and Space Administration EMU……………….. Extravehicular Mobility Unit MAG………………. Maximum Absorption Garment LCVG……………… Liquid Cooling and Ventilation Garment EEH………………... Electrical harness CCA……………….. Communications Carrier Assembly LTA………………... Lower Torso Assembly HUT………………... Hard Upper Torso IDB………………… In-suit Drink Bag SOP………………… Secondary Oxygen Pack DCM……………….. Display and Control Module SCU………………... Servicing and Cooling Umbilical AAP………………... Airlock Adapter Plate IST-I……………….. Integrated Suit Test-I POGO……………… Partial Gravity simulator MKIII……………… Mark III CTSD…………….... Crew and Thermal Systems Division HSSL…………...... Human Systems and Simulation Laboratory UC…………………. University of Cincinnati SIP…………………. Suit Interface Plate HA…………………. High-Aft HF…………………. High-Forward LF………………….. Low-Forward LA…………………. Low-Aft IRB………………… Institutional Review Board GRF………………... Ground Reaction Forces EMG……………….. Electromyography MVC……………….. Maximal Voluntary Contraction RPE………………… Rate of Perceived Exertion CMC……………….. Computed Muscle Control PD………………….. Proportional-Derivative ARGOS……………. Active Response Gravity Offload System JSC………………… Johnson Space Center

1 Introduction

This section provides a brief history of spacesuits and how astronauts perform extravehicular activity wearing them, a description to the challenges faced in designing spacesuits, our motivation for conducting this research study, and the need and specific aims for this research study. This section concludes with a brief thesis outline.

1.1 Background and Motivation

Exploring space and specifically celestial bodies such as the Moon or Mars by an astronaut or cosmonaut requires them to wear spacesuits for their survival and operation. Exploration can refer to walking on a planet’s surface or stepping out of their spacecraft to repair some subsystems or maintain a space station or conduct any extravehicular activity (EVA) task. EVA refers to the performance of tasks by an astronaut outside a spacecraft beyond Earth’s atmosphere. Spacesuits have been in use since the 1960’s starting with the first human space walk by Yuri Gagarin. There has been a lot of development on spacesuits since then (in Soviet/Russia and in the United States), such as the SK series, Berkut, Yastreb, Orlan (Russia’s current EVA suit) in Russia and the

Mercury suit, Gemini suits, Apollo suits and Extravehicular Mobility Unit (EMU) suit (currently being used by the US) [11]. Spacesuit design requires a precise test plan, especially while designing the suits based on human subject feedback. As spacesuits are designed for their missions, the goal of spacesuit design for Mars is to make it lighter and provide adequate mobility in the suits. The Apollo suits weighed 180lbs on Earth but weighed about 30lbs on Moon, however, they didn’t have a lot of mobility in them making it difficult to work in the suits [12]. Hence, the center of gravity (CG) of the spacesuit and the portable life support are very vital in designing a spacesuit.

As there are no custom spacesuits for every crew member, it is very important to understand the

role of spacesuit CG on EVA task performance. The new spacesuits are designed based on the old prototypes. However, some of the knowledge related to the spacesuits is lost and hence NASA has initiated a series of tests aimed at understanding human physiological and biomechanical effects of spacesuits under a variety of conditions. Some of the new spacesuit designs that are currently being researched on are the Z-suit, the Mark III, the I-suit and the Bio-suit. Our proposed study seeks to provide objective data that can be used to better define the influence of suit CG location and the limits of CG placement for spacesuit requirements and design. Ultimately, this information should promote EVA task perform and astronaut comfort.

1.2 Research Goals and Approach

This thesis contributes to the first part of a study that determines the spacesuit and portable life support system center of gravity influences on astronaut extravehicular activity task performance.

The primary objective of this thesis is to develop a methodology for determining the best center of gravity (CG) location for a combined spacesuit and portable life support system (PLSS) during extravehicular activity (EVA) task performance. There exists a need to quantify the influence of space suit center of gravity (CG) location on astronaut kinematics, exertion and work performed during extravehicular activity (EVA) tasks. Our specific aims are as follows:

Specific Aim 1: Determine differences in kinematics based on CG location. Kinematics relates to the joint angles and the ability to perform the task within the mobility constraints of a spacesuit.

Specific Aim 2: Demonstrate that certain CG locations require more exertion to complete an EVA task compared to other CG locations. Exertion relates to the force or torque, and power that is required.

Specific Aim 3: Demonstrate that certain CG locations help in improving efficiency of task performance compared to other CG locations. Efficiency relates to the joint work required to complete an EVA task.

1.3 Thesis Outline

This thesis is presented in six chapters. The first section introduces the research goals and approach. The second section provides a detailed introduction to spacesuits and some of the studies that were conducted on them related to this research in the form of a literature review. Section 3 provides a detailed experimental design focusing on the center of gravity variation fixture and the methods used to conduct the experiment and obtain the data. Section 4 presents the results relating to this research study, mainly focusing on the specific aims. Section 5 presents the discussion and conclusions of this study in detail. Section 6 presents recommendations for future work. This section is followed by a bibliography of the references used in this study and an appendix covering some of the scales used for this study and their raw results from our subjects.

2 Literature Review

This section provides a detailed overview on spacesuits and some National Aeronautics and Space

Administration (NASA) studies focusing on the factors contributing to the astronaut performance during Extravehicular activity (EVAs) in spacesuits.

2.1 Spacesuits

A spacesuit is the garment that protects an astronaut from the extreme environment of space. Space is hostile because there is minimal or no pressure, human-unbearable temperatures (such as 248 degrees Fahrenheit in the sun and -148 degrees Fahrenheit in the shade), radiation such as cosmic rays, charged particles emitted from the sun, micrometeoroids or orbital debris moving at high speeds. Hence, a spacesuit is designed to provide an astronaut with oxygen and remove carbon dioxide, a pressurized atmosphere that prevents the blood and body fluids from boiling, maintain a comfortable temperature, protect from micrometeoroids and from radiation to an extent, allows to see clearly and provides some movement that helps in performing tasks, and communicate with ground control, spacecraft, fellow astronauts. While spacesuit design has been classified theoretically into four different approaches namely, soft suits, hard-shell suits, hybrid suits and skintight suits, the modern space suit EMU (extravehicular mobility unit) has a combination of soft and hard components to provide support, mobility and comfort. Materials such as Nylon tricot,

Spandex, Urethane-coated Nylon, Dacron, Neoprene-coated Nylon, Mylar, Gortex, Kevlar and

Nomex are used in making the EMU suit. The suit has two layers of an inner cooling garment, two layers of pressure garment, and eight layers of thermal micrometeoroid garment followed by an outer layer. All the thirteen layers are stitched together to form the suit. The parts of the EMU

(shown in Figure 2.1 (a)) along with their roles are as follows [4]:

 Maximum Absorption Garment (MAG): a large disposable, absorbant diaper that

collects astronaut’s urine and feces while in the suit. This is needed because astronauts may

spend a lot of time working in their spacesuits and it is not possible to go inside the

spacecraft to use the toilet every time because of the pressurization and the depressurization

times (of the spacesuit and airlocks/spacecraft).

 Liquid Cooling and Ventilation Garment (LCVG): a “long underwear” with thin plastic

tubes. Cool water (coming from the backpack or from the spacecraft through an umbilical

cord) flowing through these tubes remove excess body heat produced by astronaut.

 EMU Electrical Harness (EEH): provides connections to the radio and allows

communications and monitors bio-instruments that reflect an astronaut’s vital signs

(respiration rate, heart rate, etc.).

 Communications Carrier Assembly (CCA): a fabric cap that contains microphones and

earphones to use with the radio for hands-free communication.

 Lower Torso Assembly (LTA): the lower half of the EMU including lower waist, pants,

knee and ankle joints, boots.

 Hard Upper Torso (HUT): hard fiberglass shell supporting the arms, torso, life-support

backpack and the control module.

 Arms: contain shoulder, upper arm and elbow joint bearings that let an astronaut move his

or her arms in many directions.

 Gloves: outer gloves with loops to tether tools and inner fine-fabric gloves for comfort.

 Helmet: made of clear, impact-resistant, polycarbonate plastic and padded in the rear for

comfort, the helmet connects to the HUT by a quick-connect ring. It also has a purge valve

that removes carbon dioxide if backup oxygen needs to be used.

 Extravehicular Visor Assembly (EVA): it fits over the helmet and protects the astronaut

from bright sunlight. It consists of a metallic gold covered visor and adjustable blinders to

filter and block sunlight, a clear, impact resistant cover for thermal and impact protection,

a TV camera and four headlamps.

 In-suit drink bag (IDB): an astronaut needs drinking water while working in the spacesuit

hence the IDB holds up to 32 ounces of water in a plastic pouch inside the HUT. A straw

positioned next to the astronaut’s mouth is connected to the pouch by a small tube.

 Primary/Portable Life Support System (PLSS): a backpack containing oxygen tanks,

carbon dioxide scrubbers/filters, electrical power, cooling water, radio equipment,

ventilating fans and warning systems. Oxygen flows into the suit behind the astronaut’s

head and out of the suit at the feet and elbows. Once inside the PLSS, the airflow enters a

charcoal cartridge that removes odors followed by the carbon dioxide scrubber. This gas is

then sent to a sublimator using a fan. The sublimator removes water vapor and channels

the gas back to the cooling-water supply. PLSS can remove carbon dioxide and supply

oxygen for about 7 hours. The rechargeable EMU battery supplies 27 Amp-hours of

electrical current when its 11 Zinc cells are connected in series.

 Secondary Oxygen Pack (SOP): an emergency oxygen supply pack that turns on and

supplies half an hour of oxygen when the suit pressure drops below 0.23 atm. It is attached

to the backpack frame below the PLSS. This supply of oxygen is sufficient for an astronaut

to return to the spacecraft.

 Display and Control Module (DCM): DCM mounts to the chest and consists of switches,

gauges, LCD displays and valves required to control the PLSS. Sometimes, a mirror

mounted to the sleeve helps the astronaut in viewing the DCM.

 Servicing and Cooling Umbilical (SCU): an umbilical cord containing tubes to provide

cooling water, electrical wires for power and tubes for oxygen to the EMU while the

astronaut is in the airlock preparing for the spacewalk.

 Airlock Adapter Plate (AAP): a frame mounted to the walls of the airlock that holds all

the EMU pieces while the astronaut is suiting up.

Figure 2.1: Extravehicular Mobility Unit (Source: NASA)

2.2 Related NASA Studies

NASA has initiated a series of tests to understand human performance and suit kinematics, conducted in multiple partial-gravity environments, under a variety of simulated lunar EVA conditions. These studies will help NASA optimize human performance in partial-gravity

environments based on the results that will provide evidence based recommendations for suit weight, mass, center of gravity (CG), pressure and suit kinematic constraints. However, in the study titled “Metabolic Costs and Biomechanics of Level Ambulation in a Planetary Suit”,

Integrated Suit Test-I (IST-I), the suited conditions of the test were conducted at a constant CG location and suit mass while varying the suit weight, pressure and suit kinematic constraints (waist rotation mobility locked vs. unlocked) to determine their individual effects on lunar ambulation.

This study relates to the suited (Mark III suit) and unsuited human performance of treadmill locomotion on the partial-gravity simulator (POGO). Furthermore, the CG effects were not assessed as the POGO system posed potential problems in two primary areas. First, the system CG would not line up with the gimbal axes of rotation and second, the system CG would differ from the subject’s CG [1].

NASA conducted a 10-km feasibility walkback test [2] to collect human performance data and produce a crew consensus in addition to understanding human performance limitations of the suit, thus optimizing spacesuit design for a targeted operational environment. The test was conducted on a treadmill with the subjected suited in a Mark III (MKIII) demonstrator EVA suit supported by POGO to generate lunar gravity environment. While addressing the question related to the location of system center of gravity, only the alignment of system CG with the gimbal axes of rotation and the difference between the system CG and the subject CG was answered. They did not study CG alterations and hence were unable to draw conclusions regarding the influence of

CG on astronaut performance. It was also concluded that the MKIII suit may perform well in lunar gravity but not in Martian gravity due to the fact that subjects rapidly approached near maximal physiologic effort while brisk walking in increased gravity.

In another study by NASA [3], to understand the human performance and suit kinematics under a variety of simulated lunar EVA conditions produced in a parabolic flight, interrelationship among suit weight, mass, CG, pressure, and crew anthropometrics and performance was studied. The main was to provide comparable data with human performance testing on POGO and provide design guidance to more reduced-gravity projects. This study was also designed to study the varied CG and mass on suited human performance, most importantly, assess how varying the suit CG affects the biomechanics and operator compensation. This was one of the primary objectives because, previously, investigators had looked into how CG affects human performance, but never in an

EVA suit. However for this test, the mass and simulated gravity level were kept constant. The design of the experiment was such that CG could be altered to three different locations at a constant mass. To perform CG testing, an attempt was made to use the MKIII suit on POGO, but POGO could not lift the added mass necessary to create significant changes in CG. It was initially assumed that a larger mass, just enough to alter the CG would be added to the mass-support rig during CG testing. Hence, to compare the CG results, four centers of gravity positions were chosen to be tested in parabolic flight. The mass-support rig was designed with large amounts of weight on lever arms to achieve the specific targeted CG’s. These positions were in the high and aft quadrant, which is the likely zone for the suit CG based on the results concerning the suit/subject/gimbal CG from suited IST testing on POGO. The CG locations were named backpack, Crew and Thermal

Systems Division (CTSD) and POGO (two locations of POGO). The first two were chosen as they represented the CG positions of conceptual designs for spacesuits (combined pressure garment and

PLSS). The POGO positions were chosen as it matched the system CG (combined CG of the reference subject, the spacesuit, and the equipment required to change CG) achieved during the

IST’s. The reference subject weighs 81.6 kg and is 1.83 m tall. The system CG varied based on

the subject as all subjects were not identical to the reference subject. Overall, it was found that kinematics and kinetics showed little difference between different CG conditions. However, varying CG conditions affected operator compensation. It was concluded that further study was needed to evaluate the interactions among lunar-gravity simulation analog, system CG, system mass, and subject characteristics such as anthropometry, strength, and fitness as there were inter- subject variations in subjective ratings at a given CG. It is seen that the CG locations studied were chosen and not an entire envelope of CG locations were tested to study CG effects on astronaut performance.

3 Experimental Design

This section describes the design and fabrication of the CG variation fixture that was used to perform the experimental tasks, explains how the experimental tasks were performed and the computational methods that determined the kinematics, kinetics and electromyography data of the subjects performing the tasks.

3.1 Center of Gravity Variation Fixture

The Center of Gravity (CG) variation fixture was designed based off of NASA’s spacesuit CG information provided to the Human Systems and Simulation Laboratory (HSSL) at the University of Cincinnati (UC). Since the information relating to the CG locations of specific components of the MKIII Heavy suit, MKIII Light suit, Z1 suit with the Suit Interface Plate (SIP) and the Z1 suit without the SIP was export controlled, another student in the HSSL that is a US citizen processed the data to obtain overall CG locations for suit configurations, which is not export controlled. In addition to the CG of the current spacesuits, a wider range of CG’s were determined based on the projected Portable Life Support System mass and position. The CG locations determined are provided in Table 3.1.

Table 3.1: NASA Spacesuit CG declassified data

Suit PLSS status Mass (kg) X (m) Y (m) MKIII Heavy PLSS Current 122.4472684 Centered -0.189709 0.266294 Lowered -0.189709 0.237502 (2in) PLSS Projected 109.7466812 Centered -0.174990 0.247705 Lowered -0.174990 0.221460 (2in) MKIII Light PLSS Current 120.9957727 Centered -0.191265 0.264672 Lowered -0.191265 0.235535 (2in) PLSS Projected 108.2951855 Centered -0.176531 0.245643

Lowered -0.176531 0.219046 (2in) Z1 with SIP PLSS Current 141.1284714 Centered -0.179800 0.264989 Lowered -0.179800 0.240009 (2in) PLSS Projected 128.4278842 Centered -0.166242 0.248975 Lowered -0.166242 0.226547 (2in) Z1 without PLSS Current 128.4278842 Centered -0.179849 0.248975 SIP Lowered -0.179849 0.221523 (2in) PLSS Projected 115.727297 Centered -0.164809 0.229445 Lowered -0.164809 0.204556 (2in)

The CG variation fixture had to be designed to accommodate all the CG locations that were determined from the NASA data. However, the suit mass properties alone were not sufficient to reproduce the suit CG influence on the astronaut wearer. A spacesuit weighs a different amount based on the acceleration due to gravity. A suit on Mars would weigh one-third when compared to the suit weight on Earth. Hence, to design the CG variation fixture, the spacesuit and subject combined CG needed to be determined. A healthy male subject’s weight and CG information

(95.55 kg) was used to determine the combined CG location of the subject and the spacesuit. The

CG location of the subject was determined using OpenSim, a multibody rigid-body computational platform developed at Stanford University, CA [5]. The combined CG location of the subject and spacesuit is provided in Table 3.2.

Table 3.2: Combined spacesuit and subject CG

Mass M (kg) X (m) Y (m) Subject 95.55 -0.09245 -0.00186 Suit PLSS status MK3 Heavy PLSS Current 217.9972684 -0.1470798 0.1487601 217.9972684 -0.1470798 0.1325879 PLSS Projected 205.2966812 -0.136574 0.1315513

205.2966812 -0.136574 0.1175213 MK3 Light PLSS Current 216.5457727 -0.1476635 0.147066 216.5457727 -0.1476635 0.1307854 PLSS Projected 203.8451855 -0.1371192 0.1296291 203.8451855 -0.1371192 0.1154992 Z1 with SIP PLSS Current 236.6784714 -0.144536 0.1572591 236.6784714 -0.144536 0.1423633 PLSS Projected 223.9778842 -0.1347621 0.1419675 223.9778842 -0.1347621 0.1291077 Z1 without SIP PLSS Current 223.9778842 -0.1425643 0.1419675 223.9778842 -0.1425643 0.1262271 PLSS Projected 211.277297 -0.1320844 0.1248374 211.277297 -0.1320844 0.1112046

The main design objective of the CG variation fixture (shown in Fig. 3.2) was to obtain any of the combined suit and subject CG locations. Hence, 8020 struts were used in the design of the variation fixture since the aluminum extrusions and attached brackets can be positioned in a multitude of configurations. The entire fixture was designed (using Solid Works) and fabricated at Protostar

Engineering, Inc. The reconfigurable fixture was built using a military backpack attached with

8020 struts and gym weights (25 lbs and a 10 lbs). The CG locations can be varied by translating the 8020 struts in both horizontal and vertical directions. From table 3.2, the minimum and maximum of the X and Y axes CG locations were determined and a matrix of CG locations were created. This matrix (provided in Table 3.3) corresponds to the target CG locations. Though these target CG locations could be obtained using the fixture (combined subject CG and fixture CG), to reduce the risk of injury to the subject while performing the tasks, the Y-axis CG locations were modified as shown in Table 3.3 and Figure 3.3.

Fig. 3.1: A subject wearing the CG Fig. 3.2: CG variation fixture variation fixture Table 3.3: Combined spacesuit and subject CG

Target CG Subject CG + Fixture CG X (m) Y (m) X (m) Y (m) -0.132084448 0.15725906 -0.148002742 0.120924592 -0.132084448 0.11120456 -0.148002742 0.0911845 -0.147663529 0.15725906 -0.132257987 0.120924592 -0.147663529 0.11120456 -0.132257987 0.0911845

0.17 0.16 0.15 0.14 0.13

Y (m) Y 0.12 Subject CG + Suit CG 0.11 Subject CG + Fixture CG 0.1 0.09 0.08 -0.15 -0.145 -0.14 -0.135 -0.13 -0.125 -0.12 X (m)

Fig 3.3: Comparison of Target CG and fixture + subject CG

The four CG locations to be obtained on the fixture are named High-Aft (HA), High-Forward (HF),

Low-Forward (LF) and Low-Aft (LA). The Solid Works view of the fixture in the various CG locations is shown in Fig 3.4. Also shown from OpenSim are the fixture CG locations respective to the subject, the combined fixture and subject CG locations and the reference axes that was used to determine all the calculations.

Fig 3.4: Four CG locations of interest

3.2 Methods

An Institutional Review Board (IRB) approved pilot study was developed and executed to provide objective data that defines the influence of suit CG location on extravehicular activity (EVA) task performance. The tasks that were intended for this study include climbing, object translation and walking. A method for estimating the joint torques, joint power and work was developed based on an inverse-dynamics simulation procedure using kinematic and kinetic data on a biomechanical model. This musculoskeletal full body model possesses 23 degrees of freedom with 92 musculotendon actuators representing 76 muscles of the lower extremities and the torso [5].

Kinematic data was recorded using BTS Bioengineering (Milan, Italy) motion capture system comprising of eight IR cameras. The ground reaction forces (GRF) were recorded using two force plates (Bertec Inc., OH, USA) at 400 Hz in tandem with the motion capture data (recorded at 100

Hz). Also, the surface electromyography (EMG) data was collected (at 1000 Hz) for 12 muscles on the lower back and the legs using EMG electrodes.

A total of 4 healthy male subjects between 23 – 26 years of age participated in the study. All the subjects were affiliated with the University of Cincinnati. Their height ranged from 40th to 60th percentile, and their weight ranged from 20th and 50th percentile as shown in Table 3.4. The subjects did not have any history of muscle injuries and were in good health while performing the experimental tasks.

Table 3.4: De-identified anthropometric measures of the subjects

Subject Age Height Height Weight Weight Percentile Percentile

(m) (in) (kg) (lbs) Height Weight

1 23 1.75 68.90 67.22 148.19 50 50

2 25 1.70 66.93 54.30 119.71 40 20

3 25 1.80 70.87 72.44 159.70 60 50

4 26 1.77 69.69 70.43 155.27 50 50

3.2.1 Subject Preparation

The four subjects performed the experiment in boxers or rolled up shorts without any clothing to cover the top of their bodies to mitigate any marker displacement errors while performing the tasks. The subjects wore boots while performing the tasks to limit the foot’s degree of freedom

(similar to a spacesuit). A rigid body can be represented in space by three non-collinear points.

Hence, a total of 48 retro-reflective markers were placed on the subject using temporary adhesive

pads to record the trajectories of the body movements while a subject performed a task. The marker set for the shank, tibia, arms, wrist and feet is symmetrical. The markers C7 (at the start of the spine) and V. Sacral (at the end of the spine) were used only for scaling and not during the task performance. The marker set as used in OpenSim on the full body model is shown in Fig. 3.5.

Fig. 3.5: Frontal and left side view of the model with the markers

Skin preparation was also performed on the lower back and the legs in advance of the experiment to place surface electromyography electrodes. The skin preparation included trimming of the hair followed by hair removal and cleaning the skin with alcohol to remove skin oils and minimize skin impedance and improve the signal to noise ratio. Three subjects preferred to remove the hair only around the skin determined for electromyography sensor placement while one subject preferred preparation of the entire legs and the lower back. Differential bipolar electromyography (EMG) electrodes (BTS Bioengineering, Milan, Italy) were placed on the rectus femoris, vastus lateralis, tibialis anterior, gastrocnemius medialis, biceps femoris captus long muscles of both the legs and both the erector spinae ileocostalis muscles on the lower back using temporary adhesive pads.

The electrode for the rectus femoris muscle was placed in the middle of the line connecting the anterior spina iliaca superior and the superior part of the patella (shown in Fig. 3.6). The electrode for the vastus lateralis muscle was placed at two-thirds on the line from the anterior spina iliaca superior to the lateral side of the patella as shown in Fig. 3.7. Fig 3.8 shows the sensor placement location for the tibialis anterior muscle which is at one third on the line between the tip of the fibula and the tip of the medial malleolus. The electrode placement for the gastrocnemius medialis is on the most prominent bulge of the muscle as shown in Fig. 3.9. The sensor for the biceps femoris muscle was placed at the mid-point on the line between the ischial tuberosity and the lateral epicondyle of the tibia as shown in Fig. 3.10. The sensor for the erector spinae ileocostalis is placed 1 finger width medial from the line from the posterior spina superior to the lowest point of the lower rib, at the level of L2 as shown in Fig. 3.11 [6]. The electrodes were attached on the skin using temporary adhesive pads (equipped with snap fit connectors). The conductance between the skin and the electrode was improved by the presence of the aqueous conducting gel present on the adhesive pads.

Fig. 3.6: Rectus femoris sensor placement Fig. 3.7: Vastus lateralis sensor placement (Source: http://seniam.org/) (Source: http://seniam.org/)

Fig. 3.8: Tibialis anterior sensor placement Fig. 3.9: Gastrocnemius medialis sensor (Source: http://seniam.org/) placement (Source: http://seniam.org/)

Fig. 3.10: Biceps femoris captus long sensor Fig. 3.11: Erector spinae ileocostalis sensor placement (Source: http://seniam.org/) placement (Source: http://seniam.org/)

3.2.2 Experimental Protocol

The motion capture system comprising of eight cameras, is calibrated along with the force plates

prior to the start of the experiment. All the subjects performed the experiment in a single session

lasting about 3 – 5 hours. Calibration is always done before the skin preparation for the subject.

Calibration is primarily performed to define the motion capture volume as shown in Figure 3.12

for the experiment which helps the cameras in focusing on that volume to track the retro-reflective

markers present on the subject. The calibration also involves zeroing the force plates. The EMG

electrodes are charged completely before the experiment begins. This complete process defines

the setting up of the motion capture system.

3.2.2.1 Anthropometric Measurement

All the subjects were asked for their heights and weights. In addition, the weight of each subject

was obtained from the force plates when a subject stood on them during the static pose

measurement as mentioned in Section 3.2.2.3. The weights of the boots used for the experiment were also collected from the force plates.

Fig. 3.12: Motion Capture Volume

3.2.2.2 Maximal Voluntary Contraction

To standardize and compare the electromyography results among the subjects, five manual muscle testing isometric tasks were performed to record the maximal voluntary contraction (MVC) data required to normalize the EMG results. Each task was performed thrice and the maximum of these data was considered as the maximum voltage produced by that muscle. Vocal reinforcement was provided during these tasks as is a general practice while performing MVC. Twelve EMG sensors recorded muscle data from twelve muscles respectively. MVC data were individually recorded for muscles located on both the legs while MVC for the lower back muscles was recorded simultaneously.

The first MVC task was aimed at maximally contracting the rectus femoris and the vastus lateralis muscles. The subject was seated upright on a chair and asked to hold his right leg in a horizontal

position to the seat. The investigator supporting the hamstrings, applied a downward force at the foot. The subject was asked to maintain the horizontal position as shown in Figure 3.13 [7]. The same procedure was repeated for the left leg.

Fig. 3.13: MVC for the rectus femoris and the Fig. 3.14: MVC for the gastrocnemius vastus lateralis muscles [7] muscle [7]

The second MVC task was performed for the tibialis anterior muscle. The subject seated on a chair was asked to raise his right leg and orient his right leg and orient his right foot rotated in the medial direction. The investigator rotated the foot in the opposite (outward) direction while the subject resisted [7]. This caused the tibialis anterior muscle to contact and the MVC data was recorded.

The same procedure was used for the left leg. For the third MVC task concentrating on the gastrocnemius muscle, the subject was asked to stand on his right foot and hold the left leg folded such that the tibial left leg is horizontal to the ground. With three fingers to stabilize, but not to support the body, the subject was asked to raise himself vertically as high as possible on his toes as shown in Fig. 3.14 in order to maximally contract the gastrocnemii. The subject was then asked

to resume a flat footed rest position. This experiment was repeated 25 times for one leg [7]. The same procedure was followed for the left leg.

The fourth MVC task involved the biceps femoris captus long muscle. The subject was asked to flex his left knee maintain leg in external rotation and the investigator resists knee flexion at the ankle using an upward and outward force as shown in Fig 3.15 [7]. The same was repeated for the right leg.

Fig. 3.15: MVC for the biceps femoris muscle [7]

The final task was to perform MVC on the erector spinae ileocostalis muscle. The subject resisted pulling himself back while the investigator applied force as shown in Fig. 3.16 [8]. MVC data for both the erector spinae muscles (left and right) were recorded simultaneously.

Fig. 3.16: MVC for the erector spinae ileocostalis muscle [8]

3.2.2.3 Static Pose Measurement

Upon completion of the MVC, the subjects were asked to stand to rest for about 10 minutes. The subjects then wore their boots and were asked to stand still on the force plates (one leg on each force plate) with arms apart as shown in Fig. 3.17. A capture for about 5 seconds was performed to obtain the static pose required for the scaling parameters for the generic model in OpenSim.

Voice commands were indicated for the start and the end of the capture. Two static pose captures were performed for data redundancy but only one of them was used for the analysis.

Fig. 3.17: Static Pose

3.2.2.4 Experimental Tasks

The subjects were provided details on how to perform the three experimental tasks. The tasks were performed with the CG variation fixture worn by the subjects. The subjects performed all the tasks in their boots and the EMG sensors recorded their muscle activity when they performed the tasks.

The tasks were performed in all the for different CG locations. Each task was repeated three times for data redundancy.

The three tasks were object translation, climbing and walking. The main objective of object translation was to pick up an object (10 lb ball) from one point and translating it to another point, similar to an extravehicular activity task performed by astronaut. The subjects started off in a squat position (with a leg on each force plate) with the object held at a set height, stood up and rotated to translate the object to another known position, as shown in Fig. 3.18. The subjects were asked to keep their arms and wrists in a stiff position to constrain the degrees of freedom of the arms and the wrists.

Fig. 3.18: Object Translation (Low-Aft CG) as performed by a subject

Climbing involved stepping on and off the force plates, similar to an astronaut climbing over an obstacle or stepping on and off a platform. Firstly the subjects were asked to step on one of the force plates (Force Plate 2) with their left leg followed by stepping on to the other force plate

(Force Plate 1) using their right leg. They then step off the force plates, left leg first followed by the right leg as shown in Fig. 3.19.

Fig. 3.19: Climbing (High-Front CG) as performed by a subject

The walking task involved the subject, walking from one wooden platform to another wooden platform while stepping on the force plates as shown in Fig. 3.20. It was ensured that the height of the force plates and the platforms were the same from the ground. The dimensions of the force plate are 23.7 in x 15.75 in x 4.25 in. and of the wooden platform are 49.6 in x 25.60 in x 6.7 in.

This task involved three gait cycles. The subjects were asked to start with their left leg first. The first gait cycle was on the wooden platform, the second on the force plates and the third gait cycle was on the second wooden platform. This task is similar to an astronaut walking in his spacesuit.

The subjects were also briefed about the Borg’s Rate of Perceived Exertion (RPE) chart and Corlett and Bishop Discomfort scale (provided in Appendix A). After performing each repetition of a task, the investigator noted the perceived exertion and discomfort inputs, if any, provided by a subject.

While the subjects performed the tasks, another researcher always stood in the vicinity of the

subject to protect them from falling or tipping over because of the moment produced by the CG variation fixture.

Fig. 3.20: Walking (High-Aft CG) as performed by a subject

3.3 Post Experimental Analysis

Upon completion of the experiment by the subjects, the motion capture data was analyzed in a comprehensive manner as discussed below.

3.3.1 Tracking the Motion Capture Data

The motion capture data was tracked for each frame using a predefined marker model set (same as the one defined in OpenSim). Each marker and the force plate were assigned a name which helped in tracking the data. It was ensured that the data is well tracked and the errors due to marker losses are kept to a minimum. Upon completion of tracking of one set of data (each repetition of a task generates one set of data), it was saved as .tdf file. This file was then run through a splining code written in MATLAB that creates a spline for any missing traces of data (between some frames of data). This MATLAB code also separated the motion capture data from the ground reaction forces

(from the force plates) and saves it as two different files, the marker trajectories stored in a .trc file

(which is an input format to OpenSim) and a .mot file that comprises of the ground reaction forces and the positions on the force plate at which the forces act. Any missing force plate data (especially

for the walking and climbing tasks) is made zero. The inverse kinematics, inverse dynamics and the computed muscle control were performed only on the section of data where the ground reaction forces were not zero, specifically from the time Force Plate 1 recorded its maximum ground reaction force to the time Force Plate 2 recorded its maximum ground reaction force.

3.3.2 OpenSim Analysis

The .trc and .mot files obtained were then used to derive the kinematics, kinetics and muscle forces in OpenSim. The OpenSim analysis procedure is as follows.

3.3.2.1 Scaling the model

The generic model available in OpenSim need not match the anthropometry of a subject and hence, scaling of the generic model is performed to match the subject as closely as possible. The static pose motion capture obtained for each subject helps in scaling the generic model. Scaling is performed by comparing the virtual markers present on the generic model to the experimental markers placed on the subject during the experiment [5]. Subject’s mass is required to scale the model precisely and this is obtained from the ground reaction forces (.mot file) of the static pose.

Once a model has been scaled, the CG variation fixture was added to the scaled model. The fixture at each CG location was added, hence generating four scaled models of interest. These models were used to run further analysis in OpenSim as the subject had performed the task with the fixture on at four different CG locations. In addition to the fixture, for the object translation task, the object needed to be added to the model to reproduce the exact experiment in OpenSim. Thus, the object was also added at the appropriate location (between the wrists) to the models that were used to simulate the object translation task.

3.3.2.2 Inverse Kinematics

Once the scaling and the model editing was completed, inverse kinematics was performed on the model with the experimental data to generate the generalized coordinate trajectories comprising of joint translations/rotations (in degrees). The inverse kinematics tool in OpenSim typically steps through each time frame of the experimental data and positions the model that best matches the coordinate data values and the experimental marker values for that respective time step. The objective function that minimizes the marker error in matching for each time frame is as follows

[5]:

2 2 min [ ∑ 푤푖(푥푖푒푥푝 − 푥푖푞) + ∑ 푤푗(푞푗푒푥푝 − 푞푗) ] 푞 푖휖푚푎푟푘푒푟푠 푗휖푢푛푝푟푒푠푐푟푖푏푒푑 푐표표푟푑푠

Where 푞 is the vector of generalized coordinates being solved for, 푤푖 is the marker weight, 푥푖푒푥푝 is the experimental position of the marker 푖, 푥푖푞 is the position of the corresponding marker on the model, 푞푗 = 푞푗푒푥푝, for prescribed coordinates (locked joints), 푤푗 is the coordinate weight, 푞푗푒푥푝 is the experimental value for coordinate 푗. The formulation, known as the weighted least squares summation problem, is the minimum of the sum of the weighted sum of the measure of square of the distances between experimental and model markers and the sum of the weighted square error between the experimental coordinates and the coordinates computed by inverse kinematics tool.

This formulation is solved using a quadratic programming solver with a convergence criterion of

0.0001 and a 1000 iteration limit.

3.3.2.3 Inverse Dynamics

Followed by Inverse Kinematics, an inverse dynamic analysis was performed on the model using the obtained kinematics and the ground reaction forces (external loads obtained from the force

plate data). The mass-dependent relationship from classical mechanics between force and acceleration is expressed as 퐹 = 푚푎, (where 퐹 represents the force, 푚 the mass and 푎 the acceleration) with the equations of motion. Inverse dynamics solves these equations to yield net forces and net torques at each joint which help in generating the movement of that joint. The output of inverse dynamics is the time histories of net joint torques and forces acting along the coordinate aces that generate the estimated accelerations from the measured experimental motion and the ground reaction forces applied.

3.3.2.4 Computed Muscle Control

The objective of computed muscle control (CMC) is to compute a set of muscle excitations that drives a dynamic musculoskeletal model to track the desired set of kinematics generated by inverse kinematics with the application of ground reaction forces (from the force plates). This is achieved by a combination of proportional-derivative (PD) control and static optimization.

Fig. 3.21: Schematic of the Computed Muscle Control Algorithm [9]

Consider the schematic of the CMC algorithm as shown in Fig. 3.21. This algorithm generates the excitations to the individual muscles (available on the model), the muscle forces, muscle fiber lengths along with the position, velocity and acceleration of each joint. The CMC algorithm works

in three steps. Firstly, a set of desired accelerations 푞̅̈ ∗ to drive the model coordinates 푞̅ towards the experimentally-derived coordinates 푞̅̅푒푥푝̅̅̅̅ are computed using the following PD control law:

∗ ̈ ∗ ̅̅̅ ̅̅̅ 푞̅̈ (푡 + 푇) = 푞̅̅푒푥푝̅̅̅̅ (푡 + 푇) + 푘푣 [푞̅̅푒푥푝̅̅̇̅̅(푡) − 푞̅̇(푡)] + 푘푝 [푞̅̅푒푥푝̅̅̅̅ (푡) − 푞̅(푡)]

̅̅̅ ̅̅̅ ∗ 푘푣 , 푘푝 are the velocity and position feedback gains respectively and 푞̅̈ (푡 + 푇) are the desired accelerations. The forces applied by the muscles to the body do not change instantaneously and hence the desired accelerations are computed typically for every 0.010 seconds. This interval is short enough to allow adequate control but long enough for the muscle forces to change. The errors between the model coordinates and the experimentally derived coordinates are driven to zero on achieving the desired accelerations. The velocity gains to drive these errors to zero in a critically damped manner can be achieved by using the following relation:

̅̅̅ ̅̅̅ 푘푣 = 2 ∗ √푘푝

Secondly, the actuator controls 푥̅ are computed that achieve the desired accelerations using static optimization methods. The final step is to conduct a forward dynamics simulation using the computed controls. These steps are repeated until the time is advanced to the end of the desired movement interval [9].

4 Results

This section presents the results of the experimental study, primarily focusing on each of the specific aims mentioned in Section 1.

4.1 Differences in kinematics

The differences in kinematics are determined based on the CG location. Kinematics relates to the joint angles and the ability to perform the task within the mobility constraints of a spacesuit.

Tables 4.1 and 4.2 presents the hip flexion and knee STS space suit mobility range [9] with the hip flexion and knee minimum and maximum joint angles obtained from OpenSim results respectively. The minimum and maximum hip flexion angles presented are the average minimum and maximum joint angles for that joint obtained from the four subjects. In addition, the hip flexion angle was obtained as an average of the left hip flexion and the right hip flexion for each subject, and similarly the knee angle was an average of the left knee and the right knee angle for each subject.

Table 4.1: Hip Flexion Joint Angle Comparison with STS Space Suit Mobility Range

Task CG STS Hip Flexion Hip Flexion STS Location Spacesuit Average Average Spacesuit Hip Flexion Minimum Maximum Hip Flexion Lower Joint Angle Joint Angle Upper Limit (Degrees) (Degrees) Limit (Degrees) (Degrees) Climbing HA 0 20.8940708 37.8973 90 LA 0 26.5275172 37.24391 90 HF 0 20.8246669 33.35707 90 LF 0 18.3915697 34.3289 90 Object HA 0 2.8198614 78.65197 90 Translation LA 0 5.9035501 83.65056 90 HF 0 -2.4730016 77.13918 90 LF 0 -4.8505591 81.86585 90 Walking HA 0 21.3415815 27.65398 90

LA 0 24.969902 32.14447 90 HF 0 16.2573162 25.41522 90 LF 0 16.3936266 25.01852 90

Table 4.2: Knee Joint Angle Comparison with STS Space Suit Mobility Range

Task CG STS Knee Knee STS Location Spacesuit Average Average Spacesuit Knee Maximum Minimum Knee Upper Joint Angle Joint Angle Lower Limit (Degrees) (Degrees) Limit (Degrees) (Degrees) Climbing HA -120 -54.0688 -14.5351 0 LA -120 -48.8027 -17.4605 0 HF -120 -50.9395 -16.8134 0 LF -120 -50.168 -16.4545 0 Object HA -120 -97.4498 -10.6269 0 Translation LA -120 -101.291 -6.6282 0 HF -120 -100.967 -9.58207 0 LF -120 -106.185 -8.3638 0 Walking HA -120 -39.3948 -15.5139 0 LA -120 -40.3934 -17.601 0 HF -120 -37.5721 -15.4254 0 LF -120 -40.3534 -14.9288 0

Fig. 4.1 (a) presents the comparison between the OpenSim hip flexion joint angle results and the

STS Spacesuit joint mobility range. The X-axis represents the hip flexion angles (in degrees) and the Y-axis represents the tasks performed at each CG location. The tasks are represented as _C for

Climbing, _OT for Object Translation and _W for Walking respectively. The blue bar represents the range of joint angles (minimum to maximum) obtained for hip flexion results in OpenSim and the green bar represents the STS suit mobility range. It can be noted that for the Object Translation task corresponding to the High Forward and the Low Forward CG’s, the minimum hip flexion joint angle is beyond the space suit mobility range (presented in red). However, as the standard

error is high (as seen from Fig. 4.1 (b)) for these CG locations, it’s beyond the scope of this study

(as no statistical validation is being performed in this study) to comment on it.

Fig. 4.1 (a): Hip Flexion Joint Angle Comparison with STS Spacesuit Joint Mobility Range

Figures 4.1 (b) and (c) present the averaged minimum and maximum hip flexion joint angle results from OpenSim. Comparing the average minimum hip flexion joint angles, the Low-Forward CG is the best for climbing and walking tasks, while the High-Forward CG the best for Object translation. The High-Forward CG also fares well with Low-Forward CG for the walking task.

Comparing the average maximum hip flexion angles, the High-Forward CG is the best for all the tasks. In addition, the Low-Forward CG also compares well with the High-Forward CG for walking.

Fig. 4.1 (b): Average Minimum Hip Flexion Joint Angle

Fig. 4.1 (c): Average Maximum Hip Flexion Joint Angle

Fig. 4.2 (a) presents the comparison between the OpenSim knee joint angle results and the STS

Spacesuit joint mobility range. The X-axis represents the knee angles (in degrees) and the Y-axis represents the tasks performed at each CG location. The tasks are represented as _C for Climbing,

_OT for Object Translation and _W for Walking respectively. The blue bar represents the range of joint angles (minimum to maximum) obtained for knee results in OpenSim and the green bar represents the STS suit mobility range. It can be noted that all the tasks fall within the mobility range of the spacesuit.

Fig. 4.2 (a): Knee Joint Angle Comparison with STS Spacesuit Joint Mobility Range

Figures 4.2 (b) and (c) present the averaged minimum and maximum knee joint angle results from

OpenSim. Comparing the average minimum knee joint angles, for climbing, the Low-Aft CG, for

Object translation, the High-Aft and the High-Forward CG for walking are the best.

Fig. 4.2 (b): Average Minimum Knee Joint Angle

Fig. 4.2 (c): Average Maximum Knee Joint Angle

However, for the maximum knee joint angles, the High-Aft CG is the best for climbing, the Low-

Aft CG for object translation and the Low-Forward CG for walking. The High-Aft CG is the worst for climbing and walking and the Low-Forward CG fares well with the Low-Forward CG for walking.

4.2 Exertion

The exertion required to complete an EVA task is determined. Exertion relates to force or torque, and power that is required to complete a task.

4.2.1 Joint Torques

The joint torques are a result of the inverse dynamics simulations in OpenSim. The joint torques were limited to the hip flexion and the knee for this study. An average of the left hip flexion and right hip flexion and similarly, left knee and right knee was calculated to present the joint torques for the hip flexion and the knee. The time histories of joint torques at the four different CG locations for each task is shown in figures 4.3 – 4.8. The joint torques are measured in Newton- meter (Nm) and time in seconds (s). The data was plotted using MATLAB. High-Aft CG results are shown in red, High-Forward in blue, Low-Aft in black and Low-Forward in green. As it can be seen from the time history results of the joint torques, it’s very difficult to justify if there’s any

CG location which is best suited for each task as the data is raw. When all the subject results are plotted, a time history plot cannot be used to justify good or bad CG locations. Hence, this raw data is further processed.

Fig. 4.3: Hip Flexion Joint Torque – Climbing

Fig. 4.4: Hip Flexion Joint Torque – Object Translation

Fig. 4.5: Hip Flexion Joint Torque – Walking

Fig. 4.6: Knee Joint Torque – Climbing

Fig. 4.7: Knee Joint Torque – Object Translation

Fig. 4.8: Knee Joint Torque – Walking

The joint torques for all the four subjects are obtained from inverse dynamics and the maximum of the hip flexion and knee torques are averaged among the subjects for each CG location (Hip flexion joint torques are the averages of both the left hip flexion and the right hip flexion joint torques, and similarly the knee joint torques are the averages of both the left knee and the right knee joint torques). The maximum was chosen to represent a worst case scenario. The averages are plotted as a bar chart in MATLAB as shown in figures 4.9 and 4.10. The standard error for each bar is shown in magenta. Standard error (and not standard deviation) is presented as the means are compared.

Fig. 4.9: Average Max Joint Torque – Hip Flexion

Fig 4.9 represents the average maximum joint torque for the hip flexion. It can be clearly seen that the Low-Aft CG fares best in comparison with the other CG locations for all the three tasks. The

Low-Forward CG location is also good for the walking task but doesn’t compare well with the

Low-Aft CG for the walking and object translation tasks. Fig 4.10 represents the average maximum joint torque for the knee joint. It can be realized that the Low-Front CG is the best for object translation and walking while the Low-Froward CG is best for the climbing task. Compared to the hip flexion joint torques, the Low-Aft CG is not preferable if just the knee joint torques are compared.

Fig. 4.10: Average Max Joint Torque – Knee

It can be seen that each joint presents a different conclusion about the best CG locations, the maximum hip flexion and knee joint torques were recorded for all the CG locations for each task.

For every task, a frequency of which CG location proved to be the best for each subject was noted

(the minimum joint torque was chosen among the maximum’s for hip flexion and knee to give the best CG location). As there are 4 subjects and two joints, the maximum frequency for a task is 8.

This frequency was now multiplied with 3 (a weighing factor) to compare the joint torque results

with other parameters such as joint power, muscle forces and work done, as discussed later. This frequency is plotted as a bubble chart, shown in Fig. 4.11, wherein the size of the bubble represents the frequency of best case result. Thus, the larger the size of a bubble, the more desirable the associated CG location is for the parameter being considered.

Fig. 4.11: Maximum Joint Torque Results

From Fig. 4.11, it can be concluded that the Low-Aft CG location is the best for climbing and the

High-Forward CG is the best for object translation and walking.

4.2.2 Joint Power

The joint power for the hip flexion and the knee joints is calculated as the product of the joint torque and the joint velocity. The Joint torques are obtained from Inverse Dynamics while the joint velocities are obtained after Computed Muscle Control. A MATLAB code was developed to resize the time histories of the joint velocities to match the time histories of the joint torques. Joint torques are obtained every 0.01 seconds while the joint velocities are obtained in intervals of 0.003 seconds of the simulation time. Firstly, the left hip flexion, right hip flexion, left knee and right knee power is calculated for each subject for all the CG locations for every task. Secondly, the hip flexion and

knee power is calculated as the average of the left hip flexion and right hip flexion and left knee and right knee respectively. The maximum joint power of these averages is then found for each subject at every CG location for each task. Finally, the average of the maximum joint power is calculated for each task performance at each CG location (average of four subject results). Fig.

4.12 (a) shows the average max joint power of hip flexion. The max joint power is measured in

Watts. The standard error is shown in magenta. It can be clearly seen that the Low-Aft CG location is the best for all the three tasks. The High-Forward CG location is the worst CG for climbing and walking while the Low-Forward CG is the worst for object translation.

Fig. 4.12 (a): Average Max Joint Power – Hip Flexion

Fig. 4.12 (b) represents the average max joint power for the knee joint. While the High-Forward

CG location best suits for the climbing task, the Low-Aft CG location is the best for object

translation and walking. However, the average max power for the High-Aft and the High-Forward

CG locations are on par with the Low-Aft CG location.

Fig. 4.12 (b): Average Max Joint Power – Knee

Similar to the bubble chart presented in Fig. 4.11, a bubble chart for the max joint power is shown in Fig 4.13. For every task, a frequency of which CG location proved to be the best for each subject was noted based on the minimum joint power. As there are 4 subjects and two joints, the maximum frequency for a task is 8. This frequency was now multiplied with 3 as previously done for the maximum joint torques. It can be concluded that the Low-Aft CG location is the best based on the maximum joint power results.

Fig. 4.13: Maximum Joint Power Results

4.2.3 Muscle Force

Muscle force refers to the exertion of a muscle while performing a task. The muscles that were chosen for this study were rectus femoris, vastus lateralis, biceps femoris captus long, tibialis anterior, gastrocnemius medialis and erector spinae ileocostalis, primarily focused on the legs and the lower back. The muscle forces, measured in Newton (N), are obtained from Computed Muscle

Control results in OpenSim. Every aforementioned muscle has a left and a right component, representing the left and the right parts of the human body. The muscle forces are obtained for every task from each subject. The muscle forces are averaged to obtain the joint muscle forces for each subject, for example, the left rectus femoris and the right rectus femoris results are averaged to obtain rectus femoris muscle force. The same procedure is repeated with every muscle to obtain the muscle forces for one subject. Now, the muscle forces are averaged among the four subjects to obtain the average muscle force in performing a task at a CG location. The maximum of these averages is recorded and plotted as follows using MATLAB.

Fig. 4.14: Average Maximum Muscle Force – Rectus Femoris

Fig. 4.14 shows the average maximum muscle force for rectus femoris. While the High-Aft CG is best for the object translation and walking tasks, the Low-Forward CG is the best for climbing.

However, the High-Aft fares on par with the Low-Forward. Fig. 4.15 presents the average maximum muscle force for vastus lateralis. While Low-Aft CG looks like a better CG for the climbing task, the High-Aft and High-Forward seem to be better for the object translation and the walking tasks. Fig. 4.16 shows the average maximum muscle forces for biceps femoris captus long. The Low-Aft CG looks the best CG for climbing and object translation. However, the Low-

Forward is the best for climbing task and the High-Forward fares equally well with the Low-Aft for object translation.

Fig. 4.15: Average Maximum Muscle Force – Vastus Lateralis

Fig. 4.16: Average Maximum Muscle Force – Biceps Femoris Captus Long

Figs. 4.17, 4.18 and 4.19 presents the average maximum muscle forces for tibialis anterior, gastrocnemius medialis and erector spinae ileocostalis respectively. For the tibialis anterior muscle, the Low-Forward CG is the best for climbing and walking tasks while the High-Forward is the best for object translation. The Low-Forward CG is also the best among the other CG locations for climbing and walking tasks for the gastrocnemius medialis muscle. However, the

High-Aft CG fares well with the Low-Forward for the climbing and is the best for object translation. From Fig. 4.19, it can be seen that the Low-Forward CG is the best for climbing and walking tasks from the erector spinae ileocostalis results, however, nothing can be told about the best CG location for the object translation task.

Fig. 4.17: Average Maximum Muscle Force – Tibialis Anterior

Fig. 4.18: Average Maximum Muscle Force – Gastrocnemius Medialis

Fig. 4.19: Average Maximum Muscle Force – Erector Spinae Ileocostalis

A bubble chart was created similar to max joint torque and max joint power to conclude about the best CG locations for each task based on the muscle force results. As there are 4 subjects and 6 muscles of interest, the frequency (the maximum radius of a bubble representing a task) is 24. It can be clearly seen that the High-Aft CG location favors the object translation task while the Low-

Forward CG location is the best for climbing and walking tasks.

Fig. 4.20: Maximum Muscle Force Results

4.3 Efficiency of task performance

The efficiency of a task performance is determined based on the joint work required by the subject to complete a task. Work is calculated by the joint torque integrated over the joint displacement.

Joint torques are obtained from Inverse Dynamics in OpenSim and joint displacements from

Computed Muscle Control. A MATLAB code was developed to resize the time histories of the joint displacements to match the time histories of the joint torques. Joint torques are obtained every

0.01 seconds while the joint displacements are obtained in intervals of 0.003 seconds of the simulation time. Another MATLAB code was developed which computes the work done by

integrating the joint torques over joint displacements. Firstly, the hip flexion and knee joint work are determined (left hip flexion and right hip flexion work is calculated and the total average work for the hip is found, similarly the total average work for the knee is found from the left knee joint and right knee joint work) for each subject. An average of the hip flexion and knee joint work is calculated from the results of all the four subjects for each task performance at every CG location.

The work is measured in Joules (J).

Fig. 4.21: Average Work – Hip Flexion

Fig. 4.21 presents the average work for hip flexion. It can be seen that the Low-Aft CG location is the best for performing climbing, object translation and walking. Here, it can also be seen that the

High-Forward CG location is the worst for task performance. Fig 4.22 shows the average work for the knee joint. While it’s difficult to conclude anything about the walking task, the High-Forward is the best for object translation and the Low-Forward for climbing.

Fig. 4.22: Average Work – Knee

A bubble plot is created as previously done for the other parameters, shown in Fig. 4.23. For every task, a frequency of which CG location proved to be the best for each subject was noted based on the minimum work. As there are 4 subjects and two joints, the maximum frequency for a task is 8.

This frequency was now multiplied with 3 as previously done for the maximum joint torques and maximum joint power. While the Low-Forward CG favors climbing, the High-Forward CG favors object translation and the Low-Aft CG is best suited for walking from the joint work results.

Fig. 4.23: Joint Work Results

4.4 Combined Results

A bubble plot is created, as shown in Fig. 4.24, based on the parameters previously studied, namely, the max joint torque, the max joint power, the maximum muscle force and the total work.

All their frequencies are added to achieve a total frequency of 96 for each task. Combining all the parameters, the Low-Forward CG is the best for climbing, the Low-Aft CG for object translation and the High-Forward CG for walking. Summing up the bubbles in each quadrant, the Low-Aft

CG appears to be the best compromise for the parameters and the tasks used in this study as shown by “All Tasks” in the plot.

Fig. 4.24: Combined Results

4.5 Rate of Perceived Exertion

The RPE and discomfort scale results as recorded by the investigator are provided in Appendix A.

The scale is provided for reference in Appendix A. Fig. 4.25 shows the average rate of perceived exertion based on the inputs from the subjects. For the climbing task, the Low-Forward was perceived as least difficult by the subjects while for the object translation task, except for the High-

Aft CG, all the other CG locations were felt to be the same. For walking, the High-Forward CG was the best based on the inputs provided by the subjects.

Fig. 4.25: Average Rate of Perceived Exertion

5 Discussions and Conclusions

This experimental study was conducted to determine the influence of spacesuit and portable life support system Center of Gravity (CG) on astronaut kinematics, exertion and efficiency during extravehicular activity task performance.

Motion capture was performed on the subjects while they performed the tasks with the CG variation fixture. The motion capture data was then used to compute the joint angles, the joint torques and muscle forces based on inverse kinematics, inverse dynamics and computed muscle control algorithms in OpenSim. Joint power and work required by a subject were further calculated in MATLAB. In addition to the motion capture data, the rate of perceived exertion (RPE) and the discomfort faced by a subject (if any) during task performance were recorded using the Borg’s

RPE scale and Corlett and Bishop’s discomfort scale respectively. Though three repetitions of data for each task were collected, only one set was analyzed.

Targeting the first specific aim of this experimental study, spacesuit CG influences astronaut kinematics during extravehicular task performance. Except for the hip flexion joint angles corresponding to the High Forward and the Low Forward CG’s for object translation, the hip flexion joint angles are within the space suit mobility range. However, it can be noted that all the tasks fall within the mobility range of the spacesuit when the knee joint angle is considered.

With regard to the second specific aim, spacesuit CG influences astronaut exertion during extravehicular task performance. From Fig. 4.11, it can be concluded that that the Low-Aft CG location is the best for climbing and the High-Forward CG is the best for object translation and walking with respect to the joint torques. From Fig. 4.13, the Low-Aft CG location appears to be the best based on the maximum joint power. Looking at the muscle forces, from Fig. 4.20, it can

be concluded that the High-Aft CG location favors the object translation task while the Low-

Forward CG location is the best for climbing and walking tasks.

For the third and final specific aim, spacesuit CG influences the efficiency of an astronaut during extravehicular activity task performance. From Fig. 4.21, it can concluded that while the Low-

Forward CG favors climbing, the High-Forward CG favors object translation and the Low-Aft CG is best suited for walking from the joint work results.

Combining all the parameters, from Fig. 4.22, it can be observed that the Low-Forward CG is the best for climbing, the Low-Aft CG for object translation and the High-Forward CG for walking.

This is true because for climbing the fixture doesn’t try to pull the subject back, however, for object translation, the Low-Aft CG helps in a way to counteract the moment during the inception of the task. As expected, the High-Forward CG is the best for walking, similar to daily walking with a back pack on. Summing up all the parameters, the Low-Aft CG appears to be the best compromise for the parameters and the tasks used in this study. In the end, the specific tasks and biomechanical parameters of highest importance need to be accounted for in determining the best CG location for future spacesuits.

6 Future Work

This experimental study will provide a tool to NASA to conduct future studies on spacesuit CG.

Our study was limited to a subject population of four and hence, our results could not be statistically validated. Our study was also limited by three tasks namely, climbing, object translation and walking. We propose a future study with a subject strength of 10-12 to statistically validate our results. In addition, a more diverse population based on the height and weight percentile (exploring all the ranges from 5th – 95th percentile) and age (30 – 50) can be selected.

Also, more tasks such as shoveling, repairing, prone and recover, etc. can be explored. The climbing task can be also performed using a fixer ladder or climbing on ramps of different elevations. The study can be extended for female astronauts as only male subjects were considered for our study, thus unable to conclude any CG location preferences for designing spacesuits for female astronauts. The study can also be extended to include additional biomechanical parameters such as maximal oxygen consumption, etc. NASA plans to further our study by conducting experiments on an Active Response Gravity Offload System (ARGOS) at Johnson Space Center

(JSC).

As it was shown that the spacesuit CG influences astronaut kinematics, exertion and efficiency during extravehicular task performance, the importance of each biomechanical parameter was not considered in this study. All the parameters were weighed equally. Weighing of the biomechanical parameters differently may influence the CG results which opens up another research problem.

Additional CG locations (the CG locations in the front of an astronaut, for example) can be explored to study CG influences.

References

1. Norcross JR, Clowers KG, Clark T, et al. Metabolic Costs and Biomechanics of Level

Ambulation in a Planetary Suit. NASA/TP-2010-216115. Washington, D.C.: NASA; 2010.

2. Norcorss JR, Lee LR, Clowers KG, et al. Feasibility of Performing a Suited 10-km

Ambulation on the Moon – Final Report of the EVA Walkback Test (EWT). NASA/TP-

2009-214796. Washington, D.C.; 2009.

3. Chappell SP, Norcorss JR, Clowers KG, Clark T, et al. Final Report of the Integrated

Parabolic Flight Test: Effects of Varying Gravity, Center of Gravity, and Mass on the

Movement Biomechanics and Operator Compensation of Ambulation and Exploration Tasks.

NASA/TP-2010-216137. Washington, D.C.: NASA; 2010.

4. Freudenrich, Ph.D., Craig. “How Space Suits Work” 14 December 2000.

HowStuffWorks.com http://science.howstuffworks.com/space-suit.htm [Cited on 10 July,

2015].

5. Delp SL, Anderson FC, Arnold AS, Loan P, Habib A, John CT, Guendelman E, Thelen DG.

OpenSim: Open-source Software to Create and Analyze Dynamic Simulations of Movement.

IEEE Transactions on Biomedical Engineering. (2007).

6. www.seniam.org [Cited on October 16, 2015].

7. Hislop HJ and Montgomery J, “Daniels and Worthingham’s Muscle Testing Techniques of

Manual Examination”, pp. 219-224, 2002.

8. Konrad P, “The ABC of EMG: A Practical Introduction to Kinesiological

Electromyography”, pp. 32, 2005.

9. http://msis.jsc.nasa.gov/sections/section14.htm [Cited on October 26, 2015].

10. Thelen, D.G. and Anderson, F.C., "Using computed muscle control to generate forward

dynamic simulations of human walking from experimental data, J. Biomech., 39(6):1107-

1115, 2006.

11. https://en.wikipedia.org/wiki/Space_suit#Models_of_historical_significance [Cited on

October 28, 2015].

12. http://mentalfloss.com/article/55823/15-secrets-space-suit-design [Cited on October 28,

2015].

Appendix A

Rate of Perceived Exertion [2]: The Borg Rating of Perceived Exertion (RPE) scale measures perceived exertion. RPE is used to gauge how much effort a person feels they must exert to perform a task, on a scale of 6 to 20 developed to correlate roughly with 1/10 heart rate.

Table A1: Rate of Perceived Exertion Scale

6 No exertion at all 7 Extremely light 8 9 Very light 10 11 Light 12 13 Somewhat hard 14 15 Hard (Heavy) 16 17 Very hard 18 19 Extremely hard 20 Maximal exertion  9 corresponds to "very light" exercise. For a healthy person, it is like walking slowly at

his or her own pace for some minutes.

 13 on the scale is "somewhat hard" exercise, but it still feels OK to continue.

 17, or "very hard", is very strenuous. A healthy person can still go on, but he or she

really has to push him- or herself. It feels very heavy, and the person is very tired.

 19 on the scale is an extremely strenuous exercise level. For most people this is the

most strenuous exercise they have ever experienced.

Table A2: RPE Data Log

Task/CG Object Translation Climbing Walking

Repetition 1 2 3 1 2 3 1 2 3

Table A3 shows the RPE experimental data as recorded by the investigator during the subject’s task performance.

Table A3: Rate of Perceived Exertion Experimental Data

Subject/Task/CG Object Translation Climbing Walking

Repetition 1 2 3 1 2 3 1 2 3

Sub 1 HA 16 16 16 14 14 14 14 14 14

LA 13 13 13 15 15 15 15 15 15

HF 13 13 13 11 11 11 8 8 8

LF 11 11 11 10 10 10 9 9 9

Sub 2 HA 13 13 13 10 10 10 8 8 8

LA 13 13 13 11 11 11 9 9 9

HF 13 13 13 11 11 11 8 8 8

LF 12 12 12 9 9 9 8 8 8

Sub 3 HA 13 13 13 11 11 11 11 11 11

LA 12 12 12 11 11 11 11 11 11

HF 12 12 12 8 8 8 8 8 8

LF 13 13 13 11 11 11 11 11 11

Sub 4 HA 13 13 13 11 11 11 11 11 11

LA 11 11 11 9 9 9 9 9 9

HF 11 11 11 11 11 11 9 9 9

LF 13 13 13 9 9 9 9 9 9

Corlett & Bishop discomfort scale [2]:

The body discomfort scale (0 to 10) by Corlett & Bishop will be used to rate discomfort on any and all portion(s) of the body.

Fig A1: Corlett and Bishop Discomfort Scale

Table A4: Corlett and Bishop Discomfort Scale Data Log

Task/CG Object Translation Climbing Walking

Repetition 1 2 3 1 2 3 1 2 3

Table A5 shows the Corlett and Bishop’s discomfort scale experimental data, if any, as recorded by the investigator during the subject’s task performance.

Table A5: Corlett and Bishop’s Discomfort Scale Experimental Data

Subject/Ta Object Translation Climbing Walking sk/CG

Repetition 1 2 3 1 2 3 1 2 3

Sub HA I8,QI3 I8,QI3 I8,QI3 AI6 AI6 AI6 D4,BI2 D4,BI2 D4,BI2 1 RI3 RI3 RI3 AI4,CI AI4,CI2 AI4,CI2 2

LA - - - AI6, AI6, AI6, AI6, AI6, A3 AI6, A3 A3 A3 A3 A3

HF GI4,AI4 GI4,AI4, GI4,AI4, AI4, AI4, AI4, AI2 AI2 AI2 DI3 DI3 DI3 DI3 DI3 DI3

LF I2 I2 I2 - - - AI2 AI2 AI2

Sub HA G2, P3, G2, P3, G2, P3, AI1, AI1, AI1, A3, I2 A3, I2 A3, I2 2 Q3, A2 Q3, A2 Q3, A2 A2, A2, A2, I1 I1 I1

LA R3, V3, R3, V3, R3, V3, A3 A3 A3 DI3, DI3, DI3, A3, T3, A3 T3, A3 T3, A3 A3, I3 A3, I3 I3

HF A3, E1 A3, E1 A3, E1 A3 A3 A3 AI3, AI3, AI3, A4,ZI1 A4,ZI1 A4,ZI1

LF DI0.5, DI0.5, A2 DI0.5, A2 A1, A1, A1, A0.5 A0.5 A0.5 A2 ZI0. ZI0. ZI0. 5 5 5

Sub HA CI3, CI3, BI3 CI3, BI3 ------3 BI3

LA CI2, CI2, BI2 CI2, BI2 CI2, CI2, CI2, CI2, CI2, CI2, BI2 BI2 BI2 BI2 BI2 BI2 BI2

HF CI3, CI3, BI3 CI3, BI3 ------BI3

LF GI4, GI4, HI4 GI4, HI4 ------HI4

Sub HA ------CI2,BI2 CI2,BI2 CI2,BI2, 4 W3,B3 W3,B3, W3,B3, I3 I3 I3

LA ------

HF I3,W3, I3,W3,V3 I3,W3,V3 CI2, CI2, CI2, - - - V3 BI2 BI2 BI2

LF W3,V3, W3,V3 W3,V3 - - - W3,V3 W3,V3 W3,V3 QI0.5 QI0.5 QI0.5 RI0.5 RI0.5 RI0.5