Neil Armstrong’s digital twin: An integrative approach for movement analysis in simulated space missions

Abstract

Introduction and Purpose

Space exploration is transitioning to a new era. NASA’s Artemis program is spearheading crewed missions to the Moon in the next decade, paving the way for eventual human exploration of Mars (Smith et al., 2020). However, space missions further afar from Earth require a shift in operational concepts towards an increasingly autonomous mission architecture due to communication delay (Belobrajdic et al., 2021). Analog space missions, such as AMADEE24, address these methodological challenges in a terrestrial simulation scenario (Preston & Dartnell, 2014). In order to imitate the biomechanical and perceptual constraints of suited operations in an analog mission, the astronauts wear space-suit simulators during extravehicular activities (EVA; Groemer et al., 2012). As EVAs account for the most physically-demanding and risky operations during space missions, the relevance for human movement science in space research is increasing. We propose an approach to integrate already available data sources to conduct comprehensive human movement analysis in planetary exploration scenarios. As a proof of concept, we aimed to identify sensitive kinematic and physiological markers for muscular fatigue in astronauts’ gait during the AMADEE24 mission conducted by the Austrian Space Forum.

Methods

The AMADEE24 Mars simulation took place from March 5^th until April 5^th 2024 in the Ararat province in Armenia, which was selected as simulation site for its geological similarities to areas on Mars. Six analog astronauts (two female, four male) lived isolated in a habitat for 21 days, conducting robotic, psychological and geoscientific experiments pertinent to future surface activities on Mars. Extensive high-resolution drone imaging of the analog region before the mission was conducted to compile a digital elevation model used for EVAs. Experiments in the field were conducted by two analog astronauts wearing the AOUDA space-suit simulator with two astronauts supporting them from the habitat using radio communication (Groemer et al., 2012). The recorded telemetry data of the suit included GPS location, heart rate, CO2- and O2-concentrations in the helmet as well as temperature and humidity measurements in the suit and primarily serves for medical monitoring of the astronauts during the EVA. Beneath the suit, the astronauts wore the inertial whole-body motion capture system Xsens^TM Awinda (Xsens Technologies B.V., Enschede, Netherlands) with 17 IMU-sensors and an 8-channel electromyography (EMG) system (Cometa myon, Barregio, Italy) for muscle activity recordings of the back, lower- and upper-extremities on the dominant side. For all experiments involving geoscientific operations by the analog astronauts, a remote-controlled robotic vehicle accompanied the astronauts in the field (Edlinger et al., 2022). The rover carried tools, samples and receiver for the motion-capture- and EMG-sensors. Additionally, the rover tracked the ambient conditions and took images of the surroundings, which are used to improve the 3D-environmental-model of the EVA region (see Figure 1). Biomechanical recordings were started remotely from the habitat. Prior to each EVA, astronauts prepared the experiment in the habitat following a controlled workflow.

Four generic geoscientific operations common for geological sample-taking were performed at the beginning and end of the EVA by the test subject. The operations were performed at four different sites, each site approximately 40 to 60 m apart. A previously designed traverse plan shown in the head-up-display of the suit’s helmet defined the traverses for the astronauts. All four geoscientific operations as well as the ambulatory pathways were recorded by the motion capture- and EMG-systems. In a preliminary analysis of gait alterations, we investigated the pre- and post-EVA traverses to identify muscular fatigue accumulated over the 3 to 4 h-EVAs.

The reprocessing of the motion capture recordings was performed using the software Xsens^TM MVN Analyze Pro 2024.0. Joint angles, center of mass (CoM) kinematics and torso sway, calculated as the roll angle of the T8-segment of the Xsens^TM human model around the sagital axis, were exported from Xsens^TM MVN software. Log dimensionless jerk (LDLJ) of the CoM was computed (Balasubramanian et al., 2011). A principal component analysis (PCA) over all analog astronauts was calculated based on the whole-body kinematics using the PManalyzer (Haid et al., 2019). The raw EMG signal was butterworth-filtered (10 Hz low pass cut off, 5^th order) and normalized to maximum-voluntary-contraction measurements conducted before EVA start. All variables of interest were sliced step-wise, time-normalized and averaged for a full gait cycle. A descriptive subject-wise comparison of pre- and post-patterns was performed. Data analysis was carried out in Spyder 5.4.3 IDE using Python 3.9.

Results

Eight EVAs were recorded with four of them being aborted due to LOS-procedures (loss of signal) before a post-measurement could be conducted. Eleven EMG-recordings were compromised or missing due to technical issues (data transmission, battery). Four astronauts completed the pre-post measurements, which will be further analyzed.

Self-reported perceived muscular fatigue post-EVA based on the Borg CR10 scale was 4.75 ± 1.5 (min: 3, max: 6) which corresponds to a Borg RPE rating of 13 (somewhat hard).

The CoM LDLJ increased in all four subjects between the pre-/post-condition on average by about 4.3% from -5.57 ± 0.29 to -5.80 ± 0.41. Additionally, mediolateral sway of the torso increased by ~25% (pre: 6.4 ± 3.8°, post: 8.0 ± 4.6°; see Figure 1). The results of the PCA indicate a consistent gait adaptation in all subjects between pre-/post-condition in PC1 and PC2. While PC1 was more pronounced in the post-condition, PC2 had a decreased contribution to the gait pattern. Peak muscle activity in left and right m. erector spinae increased in the analog astronaut with the highest self-reported fatigue post-EVA (RPE = 6) by 13 and 56% compared to pre-EVA measurement.

Discussion

Kinematic analysis revealed consistent adaptations in all four subjects between the pre- and post-EVA measurements regarding CoM kinematic smoothness and trunk stability. Additionally, one astronaut showed electromyographic indications of muscular fatigue in the lower back muscles (m. erector spinae). Consistent alterations in whole-body movement behaviour were detected in all test subjects.

Trunk sway and medio-lateral CoM displacement was found to significantly increase with hard to very hard self-reported rate of perceived exertion (RPE > 15 in Borg RPE; Qu & Yeo, 2011). CoM jerk behaviour during gait was sensitive to prolonged-EVA-induced fatigue in this study. Mohr and Federolf found a decline in movement smoothness, indicated by increased jerk (LDLJ), due to muscular fatigue for a lateral shuffle task (Mohr & Federolf, 2022). Elevated jerks relate to harder steps and foot placement, meaning higher impacts on the musculoskeletal system.

As also indicated by the findings of this study, the PCA turned out to be a sensitive tool to investigate small differences and adaptations in complex whole-body movements (Kobayasi et al., 2014). Higher components (PC5 -5.85% exp. var.) in the gait pattern can unveil sensitive relations between fallers and non-fallers (Kobayasi et al., 2014). Overall, our findings are in line with literature on fatigue-induced adaptations in gait patterns, even though environmental factors, tasks and carried loads differ from previous studies.

A high dependency on other technical systems and hardware is a main limitation of the proposed approach in the current configuration, making it prone to data transmission errors and other technical issues. Integrating sensors and electrodes into the garment of the astronauts and an interface to the on-board-data-handling system of the space-suit simulator could bring improvement. We note, that the validity of the gait analysis suffers from a small sample size. Individual modelling and predictions would benefit from repeated measurements of the respective astronauts. Additionally, the preliminary results are still subject to many uncontrolled external influences, which can alter the gait pattern and affect the results, e.g. surface conditions, slope and walking speed. The integration of additional data sources (e.g. GPS location) could improve the validity of the analysis. Nonetheless, classical gait analysis is prone to errors in this setting due to the high variability of the surface environment, which motivated the selection of more upstream parameters in this analysis (e.g. CoM mechanics, trunk stability, whole-body movements, EMG of lower back muscles).

Conclusion

Movement analysis in the field in the context of analog space missions can offer new insights in the motor control behaviour in complex working environments. The approach showed to be sensitive to capture kinematic and physiological indicators of muscular fatigue in the gait pattern of the analog astronauts. Especially kinematic parameters like CoM mechanics and trunk stability seem promising indicators to quantify fatigue-induced alterations. The necessity to quantify and assess surrounding external influences on the movement behaviour in order to address relevant research questions highlights the importance of a more integrative methodological approach in the context of movement analysis in analog space missions.

References

Balasubramanian, S., Melendez-Calderon, A., & Burdet, E. (2011). A robust and sensitive metric for quantifying movement smoothness. IEEE Transactions on Biomedical Engineering, 59(8), 2126-2136. https://doi.org/10.1109/TBME.2011.2179545

Belobrajdic, B., Melone, K., & Diaz-Artiles, A. (2021). Planetary extravehicular activity (EVA) risk mitigation strategies for long-duration space missions. npj Microgravity, 7, Article 16. https://doi.org/10.1038/s41526-021-00144-w

Edlinger, R., Föls, C., & Nüchter, A. (2022). An innovative pick-up and transport robot system for casualty evacuation. 2022 IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR), 67-73. https://doi.org/10.1109/SSRR56537.2022.10018818

Groemer, G. E., Hauth, S., Luger, U., Bickert, K., Sattler, B., Hauth, E., Föger, D., Schildhammer, D., Agerer, C., Ragoning, C., Sams, S., Kaineder, F., & Knoflach, M. (2012). The Aouda. X space suit simulator and its applications to astrobiology. Astrobiology, 12(2), 125-134. http://dx.doi.org/10.1089/ast.2011.0673

Haid, T. H., Zago, M., Promsri, A., Doix, A. C. M., & Federolf, P. A. (2019). PManalyzer: A software facilitating the study of sensorimotor control of whole-body movements. Frontiers in Neuroinformatics, 13, Article 24. https://doi.org/10.3389/fninf.2019.00024

Kobayashi, Y., Hobara, H., Matsushita, S., & Mochimaru, M. (2014). Key joint kinematic characteristics of the gait of fallers identified by principal component analysis. Journal of Biomechanics, 47(10), 2424-2429. https://doi.org/10.1016/j.jbiomech.2014.04.011

Mohr, M., & Federolf, P. (2022). Fatigue-related reductions in movement smoothness during a lateral shuffle and side-cutting movement. European Journal of Sport Science, 22(10), 1522-1531. https://doi.org/10.1080/17461391.2021.1960427

Preston, L. J., & Dartnell, L. R. (2014). Planetary habitability: Lessons learned from terrestrial analogues. International Journal of Astrobiology, 13(1), 81–98. https://doi.org/10.1017/S1473550413000396

Qu, X., & Yeo, J. C. (2011). Effects of load carriage and fatigue on gait characteristics. Journal of Biomechanics, 44(7), 1259-1263. https://doi.org/10.1016/j.jbiomech.2011.02.016

Smith, M., Craig, D., Herrmann, N., Mahoney, E., Krezel, J., McIntyre, N., & Goodliff, K. (2020). The Artemis program: An overview of nasa’s activities to return humans to the moon. 2022 IEEE Aerospace Conference (AERO), 1-10. http://dx.doi.org/10.1109/AERO47225.2020.9172323