SCH: ROAM-AI: Real-world Outcomes in Ambulation and Mobility through Artificial Intelligence

Human mobility enables people to work, learn, participate in their communities, and maintain independence throughout life. Neuromuscular disorders, neurological diseases, and aging can gradually reduce mobility, limiting where people go, how they interact with others, and how fully they participate in everyday life. Scientists can measure many aspects of mobility, including walking speed, daily activity, community travel, and lived experiences, but they still do not fully understand how these pieces fit together to describe mobility in natural environments. Existing methods measure movement, activity, participation, and environmental engagement separately rather than explaining how they interact to produce functional mobility. This Smart and Connected Health (SCH) project seeks to address that challenge by developing new methods to understand mobility as the combination of how people move, where they go, and how they participate in their communities. The research will advance scientific understanding of mobility while developing scalable methods for studying mobility across the lifespan. The project promotes the progress of science, supports future innovations in health and rehabilitation, and provides interdisciplinary research opportunities for students in biomechanics, artificial intelligence, rehabilitation science, geography, and data science. This project will investigate lifespan mobility in health and neuromuscular disease as a bioecological process emerging from interactions among human movement, environmental context, and behavioral participation. The hypothesis is that measures of movement performance, community mobility, and patient-reported independence, social participation, and well-being represent complementary manifestations of a common mobility process that can be quantified through integrated computational models. To test this hypothesis, ROAM-AI will combine wearable inertial sensing, home-based markerless motion capture, GPS-derived measures of community mobility, and patient-reported outcomes using explainable supervised and unsupervised machine learning methods. These analyses will identify latent constructs underlying mobility, physical activity, community participation, independence, and well-being; quantify relationships among biomechanical, behavioral, environmental, and person-reported measures; and evaluate the generalizability of these constructs across pediatric, adult, aging, and mobility-limiting populations. The project seeks to advance scientific understanding of mobility by clarifying the relationships among movement, community participation, and well-being across real-world settings. The resulting computational framework will support future research in biomechanics, rehabilitation, aging, and Smart Health. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.