Virtual Booth

Stroke survivors with residual paralysis face significant challenges in regaining independent mobility and performing activities of daily living. Current assistive devices, ranging from passive braces to rigid exoskeletons, lack the ability to interpret a user’s voluntary intent or to adapt dynamically to environmental hazards, leaving patients vulnerable to falls and secondary injury. We propose the Electromyography-Driven AI Assistive System for Stroke Rehabilitation (EDASR): a wearable, context-aware platform that fuses surface electromyography (sEMG) decoding, computer-vision and radar-based obstacle detection, soft-exosuit or wheelchair-integrated actuation, and seamless smartphone integration.

The EDASR system employs CNN-based classifiers to translate faint EMG signals from paretic muscles into real-time movement commands (currently modelled at <1000 ms latency), while stereo RGB-D cameras and FMCW radar generate a 3D environmental map to proactively avoid collisions (<500 ms response time). A centralized control unit fuses intent and context data to drive proportional assistance via exosuit cables or wheelchair motors. A unified mobile app delivers live telemetry, hazard alerts, and voice- or gesture-activated automations, such as medication reminders and e-commerce ordering, enhancing patient safety and autonomy.

Computer modelling based on available data demonstrated 82% EMG-classification accuracy across five gesture classes, 91% effective collision avoidance in simulated corridors, and a 9.2% increase in reach workspace among ten post-stroke participants compared to passive bracing. Virtually modelled early usability testing confirmed intuitive operation and rapid donning/doffing (<55 s). Over a 12-month proposed phased development plan, including sensor calibration, AI model training, pilot deployments, and safety certification, EDASR aims to deliver an economically viable, modular solution for upper- and lower-limb rehabilitation. By integrating voluntary intent, situational awareness, and daily-living automations, EDASR has the potential to redefine assistive care and accelerate functional recovery in stroke rehabilitation.