Development of the Crew State and Risk Model for Autonomous Biomedical Physiology State Monitoring and Prediction During Exploration Extravehicular Activities

NASA has completed hundreds of microgravity spacewalks or extravehicular activities (EVA) over multiple decades and spaceflight programs. During the Apollo program, only 14 EVAs were conducted on the surface of the moon. The Artemis program aims to return to the moon for sustained human presence necessitating a new paradigm of surface EVA operations. Current microgravity EVAs and planned Artemis surface EVAs will rely on ground operations teams with near real-time communication (i.e., up to 10-15 seconds roundtrip delay); however, future Mars exploration will require crew to conduct EVAs at a full level of autonomy. The Crew State and Risk Model (CSRM) is being developed to enable autonomous biomedical support through monitoring and predicting of crewmember physical and cognitive state during EVA.

The CSRM is a collection of seven core EVA physiology domains and are developed using empirical data modeling approaches. Each model subcomponent is developed and tested as crew member individualized models from EVA training analog data. The metabolic energy model estimates metabolic rate profiles from expected planetary EVA tasks. Levels of carbon dioxide (CO2) inspired within a suit are estimated through the inspired CO2 model. The human thermal regulation model predicts core body temperature and heat storage at < 3% error. The fatigue model estimates time and probability to physical exhaustion while EVA task performance predicts cognitive effectiveness and vigilance. Metabolic cost due to suited ambulation along an EVA path is generated using the traverse model. Hydration and nutrition status are predicted using the hydration, nutrition, and waste management model for readiness and recovery. Finally, the decompression sickness (DCS) model estimates risk and sensitivity to DCS based on mission profiles. Holistically, CSRM predicts key crew states through physiology determinants of performance between model interactions. This paper presents the architecture of the CSRM and details individual model development and proficiency in predictions of EVA physiology.