Develop a cable-driven gait perturbation system that can deliver precisely controlled, unexpected slip- and trip- like perturbations during treadmill walking, and investigate the biomechanical markers of balance loss onset, recovery response and outcomes (fall/balance recovery) in healthy young adults (followed by older adults) across different perturbation modalities.

Implementation of an attention-based model (transformer) showed superior activity recognition performance (95% accuracy) compared to models without attention (67% accuracy), demonstrating the utility and applicability of attention-based models in complex and dynamic movements, Investigated a minimal number of wearable motion capture sensors and their optimal placements on the body to inform sensor placement recommendations for use in real-world environments and scenarios