- Email: [email protected]
Human locomotion and workload for simulated lunar and Martian environments
Acta AstronauticaVol. 29, No. 8, pp. 613-620, 1993 Printed in Great Britain. All rights reserved
0094-5765/93 $6.00+0.00 Copyright© 1993Pergamon Preu Lid
HUMAN LOCOMOTION AND WORKLOAD FOR SIMULATED LUNAR AND MARTIAN ENVIRONMENTSt DAVA J. NEWMAN Man-Vehicle Laboratory, Rm 37-301, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, U.S.A. and HAROLD L. ALEXANDER Laboratory for Space Teleoperations and Robotics, Massachusetts Institute of Technology, Cambridge, MA 02139, U.S.A. (Received for publication 3 February 1993)
Abstract--Human locomotion in simulated lunar and Martian environments is investigated. A unique human-rated underwater treadmill and an adjustable ballasting harness simulate partial gravity in order to better understand how gravity determines the biomechaaics and energetics of human locomotion. This study has two research aspects, biomechanics and energetics. The fundamental biomechanics measurements are continuously recorded vertical forces as exerted by subjects of the treadmill which is instrumented with a force platform. Experimental results indicate that peak vertical force and stride frequency decrease as the gravity level is reduced. Foot contact time is independent of gravity level. Oxygen uptake measurements, VO2, constitute the energetics, or workload, data for this study. As theory predicts, locomotion energy requirements for lunar (l/6-g) and Martian (3/8-g) gravity levels are significantly less than at l-g. The observed variation in workload with gravity level is nonmonotonic, however, in over half the subject population. The hypothesis is offered that energy expenditure increases for lunar, as compared with Martian, locomotion due to the subject "wasting energy" for stability and posture control in simulated lunar gravity. Biomechanics data could influence advanced spacesnit design and planetary habitat design, while workload data will help define oxygen requirements for planetary life support systems.
l. INTRODUCTION The literature suggests that gravity plays a crucial role in human locomotion, but there is no clear understanding of this role [1,2]. Humans have evolved and developed in an Earth-normal l-g environment and every bone and muscle is aligned to function in Earth gravity[3]. A few investigators have used theory and experiments to study how locomotion varies in altered gravity. Using theoretical analysis, Margaria and Cavagna suggest that "compensating mechanisms may take place in subgravity locomotion" [4]. They conclude that a higher speed could be obtained by jumping. If subjects jump or lope during simulated partial gravity locomotion, then they alter their biomechanies for progression as compared to Earth-normal walking and running. In recent locomotion studies, He et al. [1] used a cable suspension apparatus to simulate partial gravity for running. Their analysis of the mechanics of running shows decreases in stride frequency, vertical landing velocity, and the angle of excursion of the limbs. Their data agree with results attained during earlier lunar simulation studies[5,6]. In sum, the theoretical tPaper IAF/IAA-91-561 presented at the 42nd Congress of the International Astronautical Federation, Montreal, Canada, 7-11 October 1991.
predictions of Margaria and Cavagna and the experimental partial gravity results of He et al. provide insight into the fundamental dynamics of running which is essential in understanding the interesting variations that occur in partial gravity locomotion. Apparent contradictions are seen in the literature regarding energy expenditures during simulated partial gravity activities. In the mid-1960s, some authors reported an increase in metabolic expenditures for simulated weightlessness and partial gravity[7-9]. However, the majority of authors published results stating that energy expenditures at various levels of reduced gravity show a decrease in metabolic rate for walking[10--12]. These metabolic observations were further substantiated for walking, loping and running [13-17]. Distinguishing between upper and lower body exercise resolves the apparent workload contradiction. Metabolic cost may increase for upper body activities, but a decrease in metabolic cost for locomotion in partial gravity is seen as compared to Earth-normal energy expenditures. Based on the literature reviewed, decreases in peak forces and stride frequencies are anticipated for associated reductions in gravity levels. Energy consumption is expected to decrease as gravity is reduced from 1 to 0-g during the underwater experiments.
613
614
DAVA J. NEWMAN a n d HAROLD L. ALEXANDER Table I, Ballasting loads for a 8.1 m, 74 kg subject Body segment
Lunar, 1/6-g (kg)
Martian, 3/8-g (kg)
2/3-g (kg)
9/10-g (kg)
Torso Left thigh Right thigh Left lower leg Right lower leg
7.3 1.3 1.3 0.5 0.5
16.4 2,8 2.8 1.2 1.2
29. I 5.1 5. I 2.1 2.1
43.7 7.6 7.6 3.1 3.1
The impetus for our partial gravity study lies in the fact that past research has touched on the importance of gravity in human locomotion, but there is a need to fill the void of knowledge for hypogravity, the entire range of gravity between microgravity (0-g) and l-g. Attempts to investigate lunar locomotion took place in the 1960s, but lack of interest and funding have left the past two decades with a vacuum of human performance data for partial gravity environments. Unanswered research questions need to be studied; they include: What is the naturally occurring gait for human motion in a partial gravity environment at a specified treadmill speed? What are the speeds associated with transitions from walking to loping and loping to running, and are they gravity dependent?J1,2,4]. This paper describes the underwater partial gravity simulation system which was designed for the study, the experimental protocol, and experimental results. Establishing a partial gravity biomechanics database is essential for future designs of vehicles, advanced spacesuits, and habitats. Identifying gait, transition speed, and oxygen consumption is a critical first step in the understanding of human performance in partial gravity. The mechanics and energetics of human locomotion in reduced gravitational fields are altered during lunar and Martian surface activities as compared to Earthnormal activities.
energy usage. It offers the advantages over suspension systems and parabolic flight of allowing free motion for a long time period, permitting assessment of both biomechanics and steady state workload. Parabolic flight allows only short periods of partial gravity simulation, while an inherent constraint of suspension systems is that the degrees of freedom for locomotion are limited. The major disadvantage of immersion is the viscosity of the water which produces drag and damping on human movements[21] so gross translation motion must be kept to a minimum. In this underwater study, partial gravity loads are provided by an adjustable ballasting (loading) harness which distributes ballast (lead weights) on a subject ranging from 0 to 100% of their dry body weight. The subject's body-segment masses and inertial properties (based on standard models[8] determine the amount of weight required to simulate partial gravity loading. Weights are distributed on five body regions and balanced across the mass center of each: the left lower leg, right lower leg, left thigh, right thigh, and torso. A 1.8m, 74kg subject is ballasted according to Table 1 for a Martian (3/8-g) underwater simulation. The adjustable ballast harness provides realistic loading for the entire range of hypogravity from 0 to 1-g.
2.2. Subjects and experimental protocol 2. METHODS AND EQUIPMENT
2.1. Partial gravity simulation technique Water immersion has been used extensively for over 25 years for astronaut extravehicular activity (EVA) training, and has been used for zero gravity simulation research[18,19] during the past four decades. Duddy[20] refers to Lilly's 1956 study which suggested the similarities between the condition of a body freely floating in space and a body suspended in water in terms of sensory deprivation, and isolation from the physical and mental stimuli present in normal log conditions. While water immersion has demonstrated its usefulness for training and simulation, it needs to be verified as a valid partial gravity simulation technique for analysis of motion and
Six healthy subjects, 4 male and 2 female, participated in the study. Their ages ranged from 24 to 39 years and body masses ranged from 59 to 81 kg (see Table 2). To qualify for participation, each subject passed a physical examination consisting of a general checkup, and ECG test, and a treadmill stress test to determine maximum oxygen consumption (VO2~). Additionally, all subjects were athletic, qualified SCUBA (self-contained breathing apparatus) divers holding either NAUI or PADI Open Water Diver qualifications (NAUI--National Association of Diving Instructors, PADI--Professional Association of Diving Instructors). The experimental protocol and equipment was approved by the NASA Ames Human Research Experiments Review Board (HRERB), the
Table 2. Subject database Subject Subject Subject Subject Subject Subject Subject
I, 2, 3, 4, 5, 6,
SI $2 $3 $4 $5 $6
Gender
Age (yr)
Height (m)
Mass (kg)
Weight (N)
M M M M F F
24 31 30 39 32 31
1.78 1.80 1.78 1.83 1.70 1.66
73.3 73.5 74.0 81.7 61.5 59.0
718 720 725 801 603 578
Human locomotion NASA Ames Man-Rating Review Board; and a comprehensive safety analysis is filed as Hazard Report # ARCX-01-NB01-H [22]. The subjects were all experienced treadmill runners and extremely comfortable underwater. Simulations took place at the Neutral Buoyancy Test Facility (NBTF) at NASA Ames Research Center (ARC), Moffett Field, Calif. The NBTF is 9 ft deep with an 11 ft dia [23]. Each subject participated in six experimental sessions after being fully trained on the treadmill. Subjects were trained on the device until their biomechanics and energy expenditure measurements reached a plateau, and were repeatable. The typical training period consisted of three complete, 3 h experimental sessions. Once data collection commenced, a different gravity level was simulated in each of six experimental sessions. Once session was a l-g control experiment with the subject exercising on the treadmill outside of the NBTF. The remaining five sessions took place underwater in the NBTF. Five gravity conditions were simulated by ballasting the subjects with weights. The five conditions were: 0, 1/6, 3/8, 2/3 g and approximately full body loading (90-100%). Subjects moved at three speeds: 0.5, 1.5 and 2.3 m/s, during each of the experimental sessions. Figure 1 contains an experimental protocol matrix.
2.3. Equipment and experimental apparatus A human-rated underwater treadmill was designed and fabricated for this study. Electric, pneumatic, and hydraulic power systems were all considered for the treadmill drive system, and an electric motor drive was chosen due to its high performance, affordability,
615
and cleanliness of operation. The low-torque, highspeed characteristics of pneumatic motors and the possibility of hydraulic fluid-leaks favoured the electric-motor alternative. The motor is mounted above water and isolated from the underwater treadmill by a 5 m flexible shaft. Triply redundant electrical isolation is implemented where necessary in the power system design. The treadmill is equipped with a force platform in order to provide measurements of peak ground reaction force and continuous vertical force, from which gait can be determined. In typical laboratory walkway experiments, where subjects pass over a force platform as they travel, measurements are time consuming to gather and difficult to reproduce. In our design these limitations are overcome by mounting a force platform directly under the belt of a motorized treadmill. This design allows for an unlimited number of force measurements over a full range of steady speeds[24]. The force platform consists of base frame supports, 4 WagezelleTM loads cells with mounting supports, and a suspended top plate underlying the treadmill belt. WagezelleTM loads cells were chosen for the force platform due to their water resistance, precision, and robust design. Each load cell requires a 5 V d.c. input, and its output signal is amplified and transmitted to a microcomputer for recording and analysis. See Fig. 2 for a treadmill top view and sideview illustration. All test equipment was calibrated before experimentation commenced. Natural frequency and nonlinearities are within acceptable ranges. The natural frequency of the force platform is well above the
Experiment DeMon Matrix: 6 x 6 x 3
~P'" ~L-
"""
S.bJect S S.qect S Subject 4 Subject 3
. . . . . . . .,_,_,_
Subject 2
Tm~[~ll ~
2.3 m/s
Treadmill q:med 1.5 m/s
-
Subject I
O-g
llS-g
~g
2n-g
1-g H20
~
1-g
Treadmill speed 0.5 nVs
Terr.
Gravity Level Fig. 1. Experimental protocol matrix.
616
DAVAJ. NEWMANand HAROLDL. ALEXANDER
Fig. 2, Submersible treadmill-mounted force platform design for partial gravity simulation. frequency band of the force trace signal. The first resonanc~ is extremely well damped and attenuated, and appears at a frequency over 60 Hz, while the
energy of the vertical force output, identified using a fast Fourier transform, is concentrated below 5 Hz. Static linearity was measured by applying known
100
1000
80 - " ~
y = 4.1655 + 0.13242X R = 0.99921
--1-" 800
g 60 - - f - -
--'f-- 600
o
i
A
40 "'I"--
400
20 -'-f--
200
o- T 0
100
200
300
400
500
600
700
Force Output (V)
Fig. 3. Submersible treadmill force plate vertical force calibraton.
800
Human locomotion loads from 0 to 981 N (0-100 kg) on the surface of the treadmill and measuring the load cell output. Figure 3 shows a linear vertical force calibration for the treadmill-equipped force platform. Subjects performing the experiments are outfitted with a commercial SCUBA facemask and supplied with surface air through an umbilical hose and a demand breathing regulator. Subjects were not outfitted with a spacesuit because emphasis is placed on "shirt-sleeve" performance, rather than the hardware design or degree to which the pressure suit and gloves affect performance. Ametek TM gas analysis equipment is used for oxygen uptake and % CO2 measurements. The expired air passes through a turbine flow meter, and a Tygon sample line tube is vented to the surface. The gas analysis system includes the following: • • • • •
model model model model model
S-3A/I oxygen analyzer N-22M single-cell zirconia sensor R-I flow control unit CD-3A carbon dioxide analyzer P-61B sensor.
Oxygen content is continuously and accurately monitored and displayed on the control unit. This flow unit regulates the flow of a single sample of the gas stream through the sensor. Gases containing up to 15% CO2 are measured by the analyzer. The CO2 sensor utilizes a thermal infrared source, optical filters, and a preamplifier. The signal is amplified and converted from analog to digital and then a microprocessor calculates the actual CO2 concentration. A U N I Q TM CIC Heartwatch, swimmer's model 8799, was used for heart rate measurements. This is a wireless exercise computer that senses the electrical signals generated by the heart in the same manner as an electrocardiogram (ECG), but it does not require attaching intrusive electrodes with lead wires to the subject's body. A comfortable rubber belt with electrodes is strapped on the subject's chest, and a transmitter equipped with sealed electronic circuits is snapped onto the belt. A wristwatch receiver is affixed to the ballasting harness. Heart rate measurements are stored in memory on the receiver, and later downloaded to a computer for data analysis.
617
force f(t). Initial data analysis was performed using program written for Matlab TM software and additional data analysis utiized KaleidagraphTM software, especially for graphical representation of the results. Video data was recorded and manually analyzed with the help for a computer program to encode the limb positions. The gas concentration measurements were coupled with flow rate data in order to reveal the subject's oxygen uptake, I;'O2, and rate of carbon dioxide production, I?CO 2. 3.
RESULTS
3. I. Peak vertical force stride frequency and contact time Normalized values of peak force were obtained by dividing the peak values detected by the subject's weight. Figure 4 shows the average for the 6 subjects' normalized peak force vs four different simulated gravity levels for all three treadmill speeds. The data reveal a significant reduction (P < 0.5) in peak force with decreasing gravity level at all speeds, as would be expected. Stride frequency measurements indicate a general trend toward a loping gait as gravity level is decreased. For locomotion 1.5 and 2.3 m/s, the plot of average stride frequency vs gravity (Fig. 5) depicts a nonlinear reduction in stride frequency as gravity level is decreased, while Fig. 6 shows no significant difference in foot contact time for various gravity levels. This suggests that the aerial phase (time between toe-off and ground contact of the opposite foot) is significantly longer for partial gravity locomotion because the contact time does not vary with gravity level, while the stride frequency decreases. The extended aerial phase, or reduction in number of strides per minute, is typical of a loping gait in which the subject's ground reaction force is greater than the pull of gravity and the subject essentially propels himself/herself into an aerial trajectory for a few seconds during locomotion. .........4t---V = 0.$ m/s ---,11-- V = l J n~ - .t - V=2.3 m~
2.4. Data analysis and reduction Vertical ground reaction forces are sampled at 1000 Hz during each run while oxygen and CO2 levels are sampled at 0.1 Hz. Lab View 2TM programs used with a Macintosh IIfx computer and a MacAdios II (GW Instruments, Inc.) data acquisition board acquire and record the raw force data from which other variables are derived. Force data were low-pass filtered using a second order Butterworth filter with a corner frequency of 30 Hz, and then average vertical ground reaction force profiles, f(t), were obtained over a stride cycle. The stride frequency and contact time, t¢ (the duration the support foot is in contact with the ground), are determined from the vertical
2~-
jd'J~
w
.,.| 21.5-
J. 2 (
] o_~ j
0-
0
0.2
0.4.
0.6
0.8 (g)
Gravity L e v e l
l
1.2
Fig. 4. Normalized peak force vs gravity, n = 6, error bars are SD.
DAVAJ. NEWMANand HAROLDL. ALEXANDER
618
i
J _-- V=O.Sm/s [ V = 1.5 m/s
- - 1 - - V* t.5 m/s ._.1- ~,- V = ~ m / '
80-
A V = 2.3 m/s
0.5 i-.--s
70.
0.4-
j.
60.
50.
.~
,.
40.
0.2
o
30-
J
0.2
0
0.4 0.6 0.8 G r a v i t y L e v e l (g)
1
1.2
0
Fig. 5. Mean stride frequency vs gravity, n = 6, error bars are SD.
1.2"
0.8
~
0.6 0.4
i
~[- --~ .i--..... 0
0.2
I
&-
' -it" -
0.4 0.6 0.8 Gravity Level (g)
1
1.2
Fig. 6. Contact time vs gravity, n = 6, error bars are SD.
n=6
180 B
V=2.3 m/s
100
80 0
I
1
Workload is primarily measured by oxygen uptake with heart rate taken as a secondary measure o f workload. Oxygen update, 1702, or the rate of oxygen consumed during exercise, vs gravity level for all three treadmill speeds is seen in Fig. 7. F o r all three speeds there is a reduction in oxygen uptake as gravity is decreased from the l-g level. For locomotion at 1.5 and 2.3 m/s, a continuous decrease in 1702 is seen for a continuous reduction in gravity. Heart rate results are shown in Fig. 8. Again, a reduction in workload is seen for the lunar and Martian simulations as compared to the gravity level approaching l-g. F o r locomotion at 1.5 and 2.3 m/s, a continuous decrease in heart rate is seen with decreasing gravity
'
1.4.
~
I
3.2. W o r k l o a d m e a s u r e m e n t s
- .A - V =2.3 m/s 1.6
I
0.2 0.4 0.6 0.8 G r a v i t y L e v e l (g)
Fig. 7. Oxygen uptake vs gravity, n -- 6 (unless noted), error bars are S D
_- V =03 m/s •...-m -- V = 13 m/s
[~
I
0
20-
0.2
0.4
0.6
Gravity Fig. 8. Heart rate vs gravity.
0.8
Human locomotion level. However, for locomotion at 0.5 m/s, the results indicate an increase in heart rate for the lunar simulation as compared to the Martian simulation. This increase is indicated directly in Fig. 8, and a trend towards an increase is indicated by extrapolating the results in Fig. 7. This indicates that at low speeds and low levels of gravity, proportionately more energy is expended for stability and posture control then in locomotion itself. 4.
to consuming oxygen for walking at 0.5 m/s during lunar simulations. Whereas, for locomotion at 0.5 m/s during the Martian simulation, subjects' comments reveal that 3/8-g is the "optimal, and most comfortable" partial gravity level. Further studies regarding this notion of wasted energy for locomotion in low gravity levels (between 0 and 3/8-g) need to be conducted. The g-level threshold for humans being able to locomote in a typical "terrestrial" upright posture using their legs effectively for movement needs to be defined through future experimentation.
D I S C U S S I O N
The biomechanics investigation of the hypothesis that the mechanics of locomotion may differ in partial gravity as compared to Earth-normal l-g, leaves us convinced that subjects alter their mechanics and change to a loping gait for lunar and Martian gravity conditions. Several biomechanics and workload measurements were reported in order to start establishing human performance characteristics of partial gravity environments. A difference was seen in workload as measured by energy consumption. We show that subjects consume less oxygen for partial gravity simulations as compared to approximate l-g levels. The nonlinear reduction in workload seen in Figs 8 and 9 for locomotion at 0.5 m/s is intriguing. For half of the subjects' oxygen consumption measurements, a lower oxygen uptake is seen for the Martian simulation as compared to the lunar simulation. Figure 9 depicts the workload results of these three subjects. The hypothesis is proposed that higher oxygen uptake is elicited for lunar locomotion at 0.5 m/s due to the notion of "wasted energy". Subjects may be "wasting energy" for stability and posture control in addition 0.2
619
4.1. Limitations of the simulation technique This underwater locomotion study is proposed as a simulation of locomotion in partial gravity. However, there are few constraints that limit the realism of the simulation, namely, the hydrodynamic viscosity inherent in the simulation technique and the additional mass added to the subjects for ballast. The drag and damping forces experienced by a subject moving in water could alter his/her locomotion and workload. This constraint was given much attention in the design of the partial gravity simulator, and the underwater treadmill was designed to enable the subject to move his/her limbs through the water without noticeably altering their center of gravity. If the treadmill was not present and subjects were required to translate their entire bodies through the water, we feel the hydrodynamic forces would invalidate the simulation. The ballast weights add mass to the subject, however, a realistic loading of the subject's body segments was desired. To this end, ballast was distributed on five body segments and applied as
B
m
0.18
-
m n m
0.16
-
0.14
-
0.12
-
0.1
-
m
m
|
0
0.06
J-
A w
0.08-
-
$
n
u
0.040
0.2
0.4
0.6
0.8
Gravity level
Fig. 9. Oxygen uptake for subjects 1-3 where velocity is 0.5 m/s.
620
DAVAJ. NEWMANand HAROLDL. ALEXANDER
close to the center of the mass of each segment as possible. A mathematical model which further investigates the hydrodynamics of a person moving through water is currently being written and tested [25]. 5. SUMMARY
Our experiments tested the hypothesis that the biomechanics and energetics of human locomotion in simulated lunar (1/6-g) and Martian (3/8-g) environments differ from 1-g human performance. An initial database for human performance in partial gravity environments was presented. These results should be further substantiated with future studies. We saw a change in the mechanics of progression which produced loping gait for the lunar and Martian gravity level simulations. A difference in the energetics required for locomotion was seen. Energy requirements are lower for partial gravity than for Earth-normal 1-g. At extremely low gravity levels, subjects may waste energy for stability and posture control as they are trying to stabilize themselves on the treadmill in addition to progressing on the treadmill. The gait analysis data can be used as an input for advanced spacesuit and habitat design. The energetics data should be kept in mind when life support requirements for future lunar and Martian planetary habitats and EVAs are specified. Acknowledgements--We would like to thank the enthusiastic subjects who participated in this study; the outstanding technical contribution of Dr T. McMahon, Dr B. Webbon, Dr D. Akin, Dr T. Sheridan, and Nick Groleau, Esq.; and the crucial support of P. Culbertson Jr, D. Smith, J. Carbo, J. James, D. Andrews, L. Whiteside, and Bill Dougherty. This research is supported by NASA Fellowship NGT50512 and the EVA Systems Branch, Advanced Life Support Division, N A S A Ames Research Center, Moffett Field, Calif. REFERENCES
1. G. He, R. Kram and T. A. McMahon, The mechanics of running under low gravity. J. appl. Physiol. (1991). 2. R. Margaria, Biomechanics of Energetics of Muscular Exercise. Oxford University Press, Oxford. (1976). 3. J. Boslough, Searching for the secrets of gravity. Natn Geogr. 563-583 (1989). 4. R. Margaria and G. A. Cavagna, Human locomotion in subgravity. Aerospace Med. 1141 (1964). 5. D. E. Hewes, Reduced-gravity simulators for studies of man's mobility in space and on the moon. Human Factors 11, 419-432 (1969). 6. J. L. Seminara and R. J. Shavelson, Lunar simulation. Human Factors 11, 451 462 (1969).
7. W. E. Springer, T. L. Stephens and I. Streimer, The metabolic cost of performing a specific exercise in a lowfriction environment. Aerospace Med. 34, 486-488 (1962). 8. E. C. Wortz, L. E. Browne, M. R. Gafvert, A. J. Macek, W. G. Robertson and W. H. Schreck, Study of astronaut capabilities to perform extravehicular maintenance and assembly functions in weightless conditons. NASA CR-859 (1966). 9. E. C. Wortz and E. J. Prescott, The effects of subgravity traction simulation on the energy costs of walking. Aerospace Med. 37, 1217-1222 (1966). 10. D. E. Hewes and A. A. Spady Jr, Evaluation of gravity simulation techniques for studies of man's selflocomotion in the lunar environment. NASA, TN D-2176 (1964). 11. W. G. Robertson and E. C. Wortz, The effect of lunar gravity on metabolic rates. Aerospace Med. 39, 799-805 (1968). 12. W. G. Sanborn and E. C. Wortz, Metabolic rates during lunar gravity simuation. Aerospace Med. 38, 380-382 (1967). 13. D. E. Hewes, R. L. Harris and A. A. Spady Jr, Comparative measurements of man's walking and running gaits in Earth and simulated lunar gravity. NASA, TN D-3363 (1966). 14. D. E. Hewes, Analysis of self-locomotive performance of lunar explorers based on experimental reduced gravity studies. NASA, TN D-3934 (1967). 15. W. Kuehneggar, H. P. Roth and F. C. Thiede, A study of man's physical capabilities on the Moon (III). Work Physiology Research Program Document No. NSL 65-153. Northrop Space Laboratories (1965). 16. W. Letko, D. E. Hewes and A. A. Spady Jr, The problems of man's adaptation to the lunar environment. National Council Committee on Heating, Bioacoustics, and Biomechanics, National Academy of Science. Ames Research Center, Moffett Field, Calif. (1966). 17. A. A. Spady Jr and W. D. Krasnow, Exploratory study of man's self-locomotion capabilities with a space suit in lunar gravity. NASA, TN D-2641 (1966). 18. C. W. Schilling, M. F. Werts and N. R. Schandelmeier (Eds), The Underwater Handbook. Plenum Press, New York (1976). 19. O. F. Trout and W. J. Bruchey, Water immersion reduced-gravity simulation. Human Factors 11,473-488 (1969). 20. J. H. Duddy, The simulation of weightlessness using water immersion techniques: an annotated bibliography. Human Factors 11, 507-540 (1969). 21. D. Akin, M. Bowden, D. Cousins and J. Paines, Biomechanics of extravehicular activity and neutral buoyancy simulation. NASA Report Number 9-88. MIT Space Systems Laboratory (1988). 22. D. J. Newman, NASA Ames Research Center Hazard Report #ARCX-01-NB01-H. Moffett Field, Calif. (1990). 23. B. W. Webbon, Neutral Buoyancy Test Facility standard operating procedures. NASA Ames Research Cenater Document #PN-86-7104-519-05 (1987). 24. R. Kram and A. Powell, A treadmill-mounted force platform. J. appl. Physiol 67, 1692-1698 (1989). 25. D. J. Newman and M. P. Hurst, Hydrodynamics modeling of underwater human locomotion. In progress.
0094-5765/93 $6.00+0.00 Copyright© 1993Pergamon Preu Lid
HUMAN LOCOMOTION AND WORKLOAD FOR SIMULATED LUNAR AND MARTIAN ENVIRONMENTSt DAVA J. NEWMAN Man-Vehicle Laboratory, Rm 37-301, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, U.S.A. and HAROLD L. ALEXANDER Laboratory for Space Teleoperations and Robotics, Massachusetts Institute of Technology, Cambridge, MA 02139, U.S.A. (Received for publication 3 February 1993)
Abstract--Human locomotion in simulated lunar and Martian environments is investigated. A unique human-rated underwater treadmill and an adjustable ballasting harness simulate partial gravity in order to better understand how gravity determines the biomechaaics and energetics of human locomotion. This study has two research aspects, biomechanics and energetics. The fundamental biomechanics measurements are continuously recorded vertical forces as exerted by subjects of the treadmill which is instrumented with a force platform. Experimental results indicate that peak vertical force and stride frequency decrease as the gravity level is reduced. Foot contact time is independent of gravity level. Oxygen uptake measurements, VO2, constitute the energetics, or workload, data for this study. As theory predicts, locomotion energy requirements for lunar (l/6-g) and Martian (3/8-g) gravity levels are significantly less than at l-g. The observed variation in workload with gravity level is nonmonotonic, however, in over half the subject population. The hypothesis is offered that energy expenditure increases for lunar, as compared with Martian, locomotion due to the subject "wasting energy" for stability and posture control in simulated lunar gravity. Biomechanics data could influence advanced spacesnit design and planetary habitat design, while workload data will help define oxygen requirements for planetary life support systems.
l. INTRODUCTION The literature suggests that gravity plays a crucial role in human locomotion, but there is no clear understanding of this role [1,2]. Humans have evolved and developed in an Earth-normal l-g environment and every bone and muscle is aligned to function in Earth gravity[3]. A few investigators have used theory and experiments to study how locomotion varies in altered gravity. Using theoretical analysis, Margaria and Cavagna suggest that "compensating mechanisms may take place in subgravity locomotion" [4]. They conclude that a higher speed could be obtained by jumping. If subjects jump or lope during simulated partial gravity locomotion, then they alter their biomechanies for progression as compared to Earth-normal walking and running. In recent locomotion studies, He et al. [1] used a cable suspension apparatus to simulate partial gravity for running. Their analysis of the mechanics of running shows decreases in stride frequency, vertical landing velocity, and the angle of excursion of the limbs. Their data agree with results attained during earlier lunar simulation studies[5,6]. In sum, the theoretical tPaper IAF/IAA-91-561 presented at the 42nd Congress of the International Astronautical Federation, Montreal, Canada, 7-11 October 1991.
predictions of Margaria and Cavagna and the experimental partial gravity results of He et al. provide insight into the fundamental dynamics of running which is essential in understanding the interesting variations that occur in partial gravity locomotion. Apparent contradictions are seen in the literature regarding energy expenditures during simulated partial gravity activities. In the mid-1960s, some authors reported an increase in metabolic expenditures for simulated weightlessness and partial gravity[7-9]. However, the majority of authors published results stating that energy expenditures at various levels of reduced gravity show a decrease in metabolic rate for walking[10--12]. These metabolic observations were further substantiated for walking, loping and running [13-17]. Distinguishing between upper and lower body exercise resolves the apparent workload contradiction. Metabolic cost may increase for upper body activities, but a decrease in metabolic cost for locomotion in partial gravity is seen as compared to Earth-normal energy expenditures. Based on the literature reviewed, decreases in peak forces and stride frequencies are anticipated for associated reductions in gravity levels. Energy consumption is expected to decrease as gravity is reduced from 1 to 0-g during the underwater experiments.
613
614
DAVA J. NEWMAN a n d HAROLD L. ALEXANDER Table I, Ballasting loads for a 8.1 m, 74 kg subject Body segment
Lunar, 1/6-g (kg)
Martian, 3/8-g (kg)
2/3-g (kg)
9/10-g (kg)
Torso Left thigh Right thigh Left lower leg Right lower leg
7.3 1.3 1.3 0.5 0.5
16.4 2,8 2.8 1.2 1.2
29. I 5.1 5. I 2.1 2.1
43.7 7.6 7.6 3.1 3.1
The impetus for our partial gravity study lies in the fact that past research has touched on the importance of gravity in human locomotion, but there is a need to fill the void of knowledge for hypogravity, the entire range of gravity between microgravity (0-g) and l-g. Attempts to investigate lunar locomotion took place in the 1960s, but lack of interest and funding have left the past two decades with a vacuum of human performance data for partial gravity environments. Unanswered research questions need to be studied; they include: What is the naturally occurring gait for human motion in a partial gravity environment at a specified treadmill speed? What are the speeds associated with transitions from walking to loping and loping to running, and are they gravity dependent?J1,2,4]. This paper describes the underwater partial gravity simulation system which was designed for the study, the experimental protocol, and experimental results. Establishing a partial gravity biomechanics database is essential for future designs of vehicles, advanced spacesuits, and habitats. Identifying gait, transition speed, and oxygen consumption is a critical first step in the understanding of human performance in partial gravity. The mechanics and energetics of human locomotion in reduced gravitational fields are altered during lunar and Martian surface activities as compared to Earthnormal activities.
energy usage. It offers the advantages over suspension systems and parabolic flight of allowing free motion for a long time period, permitting assessment of both biomechanics and steady state workload. Parabolic flight allows only short periods of partial gravity simulation, while an inherent constraint of suspension systems is that the degrees of freedom for locomotion are limited. The major disadvantage of immersion is the viscosity of the water which produces drag and damping on human movements[21] so gross translation motion must be kept to a minimum. In this underwater study, partial gravity loads are provided by an adjustable ballasting (loading) harness which distributes ballast (lead weights) on a subject ranging from 0 to 100% of their dry body weight. The subject's body-segment masses and inertial properties (based on standard models[8] determine the amount of weight required to simulate partial gravity loading. Weights are distributed on five body regions and balanced across the mass center of each: the left lower leg, right lower leg, left thigh, right thigh, and torso. A 1.8m, 74kg subject is ballasted according to Table 1 for a Martian (3/8-g) underwater simulation. The adjustable ballast harness provides realistic loading for the entire range of hypogravity from 0 to 1-g.
2.2. Subjects and experimental protocol 2. METHODS AND EQUIPMENT
2.1. Partial gravity simulation technique Water immersion has been used extensively for over 25 years for astronaut extravehicular activity (EVA) training, and has been used for zero gravity simulation research[18,19] during the past four decades. Duddy[20] refers to Lilly's 1956 study which suggested the similarities between the condition of a body freely floating in space and a body suspended in water in terms of sensory deprivation, and isolation from the physical and mental stimuli present in normal log conditions. While water immersion has demonstrated its usefulness for training and simulation, it needs to be verified as a valid partial gravity simulation technique for analysis of motion and
Six healthy subjects, 4 male and 2 female, participated in the study. Their ages ranged from 24 to 39 years and body masses ranged from 59 to 81 kg (see Table 2). To qualify for participation, each subject passed a physical examination consisting of a general checkup, and ECG test, and a treadmill stress test to determine maximum oxygen consumption (VO2~). Additionally, all subjects were athletic, qualified SCUBA (self-contained breathing apparatus) divers holding either NAUI or PADI Open Water Diver qualifications (NAUI--National Association of Diving Instructors, PADI--Professional Association of Diving Instructors). The experimental protocol and equipment was approved by the NASA Ames Human Research Experiments Review Board (HRERB), the
Table 2. Subject database Subject Subject Subject Subject Subject Subject Subject
I, 2, 3, 4, 5, 6,
SI $2 $3 $4 $5 $6
Gender
Age (yr)
Height (m)
Mass (kg)
Weight (N)
M M M M F F
24 31 30 39 32 31
1.78 1.80 1.78 1.83 1.70 1.66
73.3 73.5 74.0 81.7 61.5 59.0
718 720 725 801 603 578
Human locomotion NASA Ames Man-Rating Review Board; and a comprehensive safety analysis is filed as Hazard Report # ARCX-01-NB01-H [22]. The subjects were all experienced treadmill runners and extremely comfortable underwater. Simulations took place at the Neutral Buoyancy Test Facility (NBTF) at NASA Ames Research Center (ARC), Moffett Field, Calif. The NBTF is 9 ft deep with an 11 ft dia [23]. Each subject participated in six experimental sessions after being fully trained on the treadmill. Subjects were trained on the device until their biomechanics and energy expenditure measurements reached a plateau, and were repeatable. The typical training period consisted of three complete, 3 h experimental sessions. Once data collection commenced, a different gravity level was simulated in each of six experimental sessions. Once session was a l-g control experiment with the subject exercising on the treadmill outside of the NBTF. The remaining five sessions took place underwater in the NBTF. Five gravity conditions were simulated by ballasting the subjects with weights. The five conditions were: 0, 1/6, 3/8, 2/3 g and approximately full body loading (90-100%). Subjects moved at three speeds: 0.5, 1.5 and 2.3 m/s, during each of the experimental sessions. Figure 1 contains an experimental protocol matrix.
2.3. Equipment and experimental apparatus A human-rated underwater treadmill was designed and fabricated for this study. Electric, pneumatic, and hydraulic power systems were all considered for the treadmill drive system, and an electric motor drive was chosen due to its high performance, affordability,
615
and cleanliness of operation. The low-torque, highspeed characteristics of pneumatic motors and the possibility of hydraulic fluid-leaks favoured the electric-motor alternative. The motor is mounted above water and isolated from the underwater treadmill by a 5 m flexible shaft. Triply redundant electrical isolation is implemented where necessary in the power system design. The treadmill is equipped with a force platform in order to provide measurements of peak ground reaction force and continuous vertical force, from which gait can be determined. In typical laboratory walkway experiments, where subjects pass over a force platform as they travel, measurements are time consuming to gather and difficult to reproduce. In our design these limitations are overcome by mounting a force platform directly under the belt of a motorized treadmill. This design allows for an unlimited number of force measurements over a full range of steady speeds[24]. The force platform consists of base frame supports, 4 WagezelleTM loads cells with mounting supports, and a suspended top plate underlying the treadmill belt. WagezelleTM loads cells were chosen for the force platform due to their water resistance, precision, and robust design. Each load cell requires a 5 V d.c. input, and its output signal is amplified and transmitted to a microcomputer for recording and analysis. See Fig. 2 for a treadmill top view and sideview illustration. All test equipment was calibrated before experimentation commenced. Natural frequency and nonlinearities are within acceptable ranges. The natural frequency of the force platform is well above the
Experiment DeMon Matrix: 6 x 6 x 3
~P'" ~L-
"""
S.bJect S S.qect S Subject 4 Subject 3
. . . . . . . .,_,_,_
Subject 2
Tm~[~ll ~
2.3 m/s
Treadmill q:med 1.5 m/s
-
Subject I
O-g
llS-g
~g
2n-g
1-g H20
~
1-g
Treadmill speed 0.5 nVs
Terr.
Gravity Level Fig. 1. Experimental protocol matrix.
616
DAVAJ. NEWMANand HAROLDL. ALEXANDER
Fig. 2, Submersible treadmill-mounted force platform design for partial gravity simulation. frequency band of the force trace signal. The first resonanc~ is extremely well damped and attenuated, and appears at a frequency over 60 Hz, while the
energy of the vertical force output, identified using a fast Fourier transform, is concentrated below 5 Hz. Static linearity was measured by applying known
100
1000
80 - " ~
y = 4.1655 + 0.13242X R = 0.99921
--1-" 800
g 60 - - f - -
--'f-- 600
o
i
A
40 "'I"--
400
20 -'-f--
200
o- T 0
100
200
300
400
500
600
700
Force Output (V)
Fig. 3. Submersible treadmill force plate vertical force calibraton.
800
Human locomotion loads from 0 to 981 N (0-100 kg) on the surface of the treadmill and measuring the load cell output. Figure 3 shows a linear vertical force calibration for the treadmill-equipped force platform. Subjects performing the experiments are outfitted with a commercial SCUBA facemask and supplied with surface air through an umbilical hose and a demand breathing regulator. Subjects were not outfitted with a spacesuit because emphasis is placed on "shirt-sleeve" performance, rather than the hardware design or degree to which the pressure suit and gloves affect performance. Ametek TM gas analysis equipment is used for oxygen uptake and % CO2 measurements. The expired air passes through a turbine flow meter, and a Tygon sample line tube is vented to the surface. The gas analysis system includes the following: • • • • •
model model model model model
S-3A/I oxygen analyzer N-22M single-cell zirconia sensor R-I flow control unit CD-3A carbon dioxide analyzer P-61B sensor.
Oxygen content is continuously and accurately monitored and displayed on the control unit. This flow unit regulates the flow of a single sample of the gas stream through the sensor. Gases containing up to 15% CO2 are measured by the analyzer. The CO2 sensor utilizes a thermal infrared source, optical filters, and a preamplifier. The signal is amplified and converted from analog to digital and then a microprocessor calculates the actual CO2 concentration. A U N I Q TM CIC Heartwatch, swimmer's model 8799, was used for heart rate measurements. This is a wireless exercise computer that senses the electrical signals generated by the heart in the same manner as an electrocardiogram (ECG), but it does not require attaching intrusive electrodes with lead wires to the subject's body. A comfortable rubber belt with electrodes is strapped on the subject's chest, and a transmitter equipped with sealed electronic circuits is snapped onto the belt. A wristwatch receiver is affixed to the ballasting harness. Heart rate measurements are stored in memory on the receiver, and later downloaded to a computer for data analysis.
617
force f(t). Initial data analysis was performed using program written for Matlab TM software and additional data analysis utiized KaleidagraphTM software, especially for graphical representation of the results. Video data was recorded and manually analyzed with the help for a computer program to encode the limb positions. The gas concentration measurements were coupled with flow rate data in order to reveal the subject's oxygen uptake, I;'O2, and rate of carbon dioxide production, I?CO 2. 3.
RESULTS
3. I. Peak vertical force stride frequency and contact time Normalized values of peak force were obtained by dividing the peak values detected by the subject's weight. Figure 4 shows the average for the 6 subjects' normalized peak force vs four different simulated gravity levels for all three treadmill speeds. The data reveal a significant reduction (P < 0.5) in peak force with decreasing gravity level at all speeds, as would be expected. Stride frequency measurements indicate a general trend toward a loping gait as gravity level is decreased. For locomotion 1.5 and 2.3 m/s, the plot of average stride frequency vs gravity (Fig. 5) depicts a nonlinear reduction in stride frequency as gravity level is decreased, while Fig. 6 shows no significant difference in foot contact time for various gravity levels. This suggests that the aerial phase (time between toe-off and ground contact of the opposite foot) is significantly longer for partial gravity locomotion because the contact time does not vary with gravity level, while the stride frequency decreases. The extended aerial phase, or reduction in number of strides per minute, is typical of a loping gait in which the subject's ground reaction force is greater than the pull of gravity and the subject essentially propels himself/herself into an aerial trajectory for a few seconds during locomotion. .........4t---V = 0.$ m/s ---,11-- V = l J n~ - .t - V=2.3 m~
2.4. Data analysis and reduction Vertical ground reaction forces are sampled at 1000 Hz during each run while oxygen and CO2 levels are sampled at 0.1 Hz. Lab View 2TM programs used with a Macintosh IIfx computer and a MacAdios II (GW Instruments, Inc.) data acquisition board acquire and record the raw force data from which other variables are derived. Force data were low-pass filtered using a second order Butterworth filter with a corner frequency of 30 Hz, and then average vertical ground reaction force profiles, f(t), were obtained over a stride cycle. The stride frequency and contact time, t¢ (the duration the support foot is in contact with the ground), are determined from the vertical
2~-
jd'J~
w
.,.| 21.5-
J. 2 (
] o_~ j
0-
0
0.2
0.4.
0.6
0.8 (g)
Gravity L e v e l
l
1.2
Fig. 4. Normalized peak force vs gravity, n = 6, error bars are SD.
DAVAJ. NEWMANand HAROLDL. ALEXANDER
618
i
J _-- V=O.Sm/s [ V = 1.5 m/s
- - 1 - - V* t.5 m/s ._.1- ~,- V = ~ m / '
80-
A V = 2.3 m/s
0.5 i-.--s
70.
0.4-
j.
60.
50.
.~
,.
40.
0.2
o
30-
J
0.2
0
0.4 0.6 0.8 G r a v i t y L e v e l (g)
1
1.2
0
Fig. 5. Mean stride frequency vs gravity, n = 6, error bars are SD.
1.2"
0.8
~
0.6 0.4
i
~[- --~ .i--..... 0
0.2
I
&-
' -it" -
0.4 0.6 0.8 Gravity Level (g)
1
1.2
Fig. 6. Contact time vs gravity, n = 6, error bars are SD.
n=6
180 B
V=2.3 m/s
100
80 0
I
1
Workload is primarily measured by oxygen uptake with heart rate taken as a secondary measure o f workload. Oxygen update, 1702, or the rate of oxygen consumed during exercise, vs gravity level for all three treadmill speeds is seen in Fig. 7. F o r all three speeds there is a reduction in oxygen uptake as gravity is decreased from the l-g level. For locomotion at 1.5 and 2.3 m/s, a continuous decrease in 1702 is seen for a continuous reduction in gravity. Heart rate results are shown in Fig. 8. Again, a reduction in workload is seen for the lunar and Martian simulations as compared to the gravity level approaching l-g. F o r locomotion at 1.5 and 2.3 m/s, a continuous decrease in heart rate is seen with decreasing gravity
'
1.4.
~
I
3.2. W o r k l o a d m e a s u r e m e n t s
- .A - V =2.3 m/s 1.6
I
0.2 0.4 0.6 0.8 G r a v i t y L e v e l (g)
Fig. 7. Oxygen uptake vs gravity, n -- 6 (unless noted), error bars are S D
_- V =03 m/s •...-m -- V = 13 m/s
[~
I
0
20-
0.2
0.4
0.6
Gravity Fig. 8. Heart rate vs gravity.
0.8
Human locomotion level. However, for locomotion at 0.5 m/s, the results indicate an increase in heart rate for the lunar simulation as compared to the Martian simulation. This increase is indicated directly in Fig. 8, and a trend towards an increase is indicated by extrapolating the results in Fig. 7. This indicates that at low speeds and low levels of gravity, proportionately more energy is expended for stability and posture control then in locomotion itself. 4.
to consuming oxygen for walking at 0.5 m/s during lunar simulations. Whereas, for locomotion at 0.5 m/s during the Martian simulation, subjects' comments reveal that 3/8-g is the "optimal, and most comfortable" partial gravity level. Further studies regarding this notion of wasted energy for locomotion in low gravity levels (between 0 and 3/8-g) need to be conducted. The g-level threshold for humans being able to locomote in a typical "terrestrial" upright posture using their legs effectively for movement needs to be defined through future experimentation.
D I S C U S S I O N
The biomechanics investigation of the hypothesis that the mechanics of locomotion may differ in partial gravity as compared to Earth-normal l-g, leaves us convinced that subjects alter their mechanics and change to a loping gait for lunar and Martian gravity conditions. Several biomechanics and workload measurements were reported in order to start establishing human performance characteristics of partial gravity environments. A difference was seen in workload as measured by energy consumption. We show that subjects consume less oxygen for partial gravity simulations as compared to approximate l-g levels. The nonlinear reduction in workload seen in Figs 8 and 9 for locomotion at 0.5 m/s is intriguing. For half of the subjects' oxygen consumption measurements, a lower oxygen uptake is seen for the Martian simulation as compared to the lunar simulation. Figure 9 depicts the workload results of these three subjects. The hypothesis is proposed that higher oxygen uptake is elicited for lunar locomotion at 0.5 m/s due to the notion of "wasted energy". Subjects may be "wasting energy" for stability and posture control in addition 0.2
619
4.1. Limitations of the simulation technique This underwater locomotion study is proposed as a simulation of locomotion in partial gravity. However, there are few constraints that limit the realism of the simulation, namely, the hydrodynamic viscosity inherent in the simulation technique and the additional mass added to the subjects for ballast. The drag and damping forces experienced by a subject moving in water could alter his/her locomotion and workload. This constraint was given much attention in the design of the partial gravity simulator, and the underwater treadmill was designed to enable the subject to move his/her limbs through the water without noticeably altering their center of gravity. If the treadmill was not present and subjects were required to translate their entire bodies through the water, we feel the hydrodynamic forces would invalidate the simulation. The ballast weights add mass to the subject, however, a realistic loading of the subject's body segments was desired. To this end, ballast was distributed on five body segments and applied as
B
m
0.18
-
m n m
0.16
-
0.14
-
0.12
-
0.1
-
m
m
|
0
0.06
J-
A w
0.08-
-
$
n
u
0.040
0.2
0.4
0.6
0.8
Gravity level
Fig. 9. Oxygen uptake for subjects 1-3 where velocity is 0.5 m/s.
620
DAVAJ. NEWMANand HAROLDL. ALEXANDER
close to the center of the mass of each segment as possible. A mathematical model which further investigates the hydrodynamics of a person moving through water is currently being written and tested [25]. 5. SUMMARY
Our experiments tested the hypothesis that the biomechanics and energetics of human locomotion in simulated lunar (1/6-g) and Martian (3/8-g) environments differ from 1-g human performance. An initial database for human performance in partial gravity environments was presented. These results should be further substantiated with future studies. We saw a change in the mechanics of progression which produced loping gait for the lunar and Martian gravity level simulations. A difference in the energetics required for locomotion was seen. Energy requirements are lower for partial gravity than for Earth-normal 1-g. At extremely low gravity levels, subjects may waste energy for stability and posture control as they are trying to stabilize themselves on the treadmill in addition to progressing on the treadmill. The gait analysis data can be used as an input for advanced spacesuit and habitat design. The energetics data should be kept in mind when life support requirements for future lunar and Martian planetary habitats and EVAs are specified. Acknowledgements--We would like to thank the enthusiastic subjects who participated in this study; the outstanding technical contribution of Dr T. McMahon, Dr B. Webbon, Dr D. Akin, Dr T. Sheridan, and Nick Groleau, Esq.; and the crucial support of P. Culbertson Jr, D. Smith, J. Carbo, J. James, D. Andrews, L. Whiteside, and Bill Dougherty. This research is supported by NASA Fellowship NGT50512 and the EVA Systems Branch, Advanced Life Support Division, N A S A Ames Research Center, Moffett Field, Calif. REFERENCES
1. G. He, R. Kram and T. A. McMahon, The mechanics of running under low gravity. J. appl. Physiol. (1991). 2. R. Margaria, Biomechanics of Energetics of Muscular Exercise. Oxford University Press, Oxford. (1976). 3. J. Boslough, Searching for the secrets of gravity. Natn Geogr. 563-583 (1989). 4. R. Margaria and G. A. Cavagna, Human locomotion in subgravity. Aerospace Med. 1141 (1964). 5. D. E. Hewes, Reduced-gravity simulators for studies of man's mobility in space and on the moon. Human Factors 11, 419-432 (1969). 6. J. L. Seminara and R. J. Shavelson, Lunar simulation. Human Factors 11, 451 462 (1969).
7. W. E. Springer, T. L. Stephens and I. Streimer, The metabolic cost of performing a specific exercise in a lowfriction environment. Aerospace Med. 34, 486-488 (1962). 8. E. C. Wortz, L. E. Browne, M. R. Gafvert, A. J. Macek, W. G. Robertson and W. H. Schreck, Study of astronaut capabilities to perform extravehicular maintenance and assembly functions in weightless conditons. NASA CR-859 (1966). 9. E. C. Wortz and E. J. Prescott, The effects of subgravity traction simulation on the energy costs of walking. Aerospace Med. 37, 1217-1222 (1966). 10. D. E. Hewes and A. A. Spady Jr, Evaluation of gravity simulation techniques for studies of man's selflocomotion in the lunar environment. NASA, TN D-2176 (1964). 11. W. G. Robertson and E. C. Wortz, The effect of lunar gravity on metabolic rates. Aerospace Med. 39, 799-805 (1968). 12. W. G. Sanborn and E. C. Wortz, Metabolic rates during lunar gravity simuation. Aerospace Med. 38, 380-382 (1967). 13. D. E. Hewes, R. L. Harris and A. A. Spady Jr, Comparative measurements of man's walking and running gaits in Earth and simulated lunar gravity. NASA, TN D-3363 (1966). 14. D. E. Hewes, Analysis of self-locomotive performance of lunar explorers based on experimental reduced gravity studies. NASA, TN D-3934 (1967). 15. W. Kuehneggar, H. P. Roth and F. C. Thiede, A study of man's physical capabilities on the Moon (III). Work Physiology Research Program Document No. NSL 65-153. Northrop Space Laboratories (1965). 16. W. Letko, D. E. Hewes and A. A. Spady Jr, The problems of man's adaptation to the lunar environment. National Council Committee on Heating, Bioacoustics, and Biomechanics, National Academy of Science. Ames Research Center, Moffett Field, Calif. (1966). 17. A. A. Spady Jr and W. D. Krasnow, Exploratory study of man's self-locomotion capabilities with a space suit in lunar gravity. NASA, TN D-2641 (1966). 18. C. W. Schilling, M. F. Werts and N. R. Schandelmeier (Eds), The Underwater Handbook. Plenum Press, New York (1976). 19. O. F. Trout and W. J. Bruchey, Water immersion reduced-gravity simulation. Human Factors 11,473-488 (1969). 20. J. H. Duddy, The simulation of weightlessness using water immersion techniques: an annotated bibliography. Human Factors 11, 507-540 (1969). 21. D. Akin, M. Bowden, D. Cousins and J. Paines, Biomechanics of extravehicular activity and neutral buoyancy simulation. NASA Report Number 9-88. MIT Space Systems Laboratory (1988). 22. D. J. Newman, NASA Ames Research Center Hazard Report #ARCX-01-NB01-H. Moffett Field, Calif. (1990). 23. B. W. Webbon, Neutral Buoyancy Test Facility standard operating procedures. NASA Ames Research Cenater Document #PN-86-7104-519-05 (1987). 24. R. Kram and A. Powell, A treadmill-mounted force platform. J. appl. Physiol 67, 1692-1698 (1989). 25. D. J. Newman and M. P. Hurst, Hydrodynamics modeling of underwater human locomotion. In progress.