Adaptive Human-in-the loop Control of Soft Exoskeleton Using Wearable Cardio-Respiratory Sensor

The decline in walking efficiency associated with aging and physical impairment represents a major challenge for mobility and independence. Soft wearable exosuits have emerged as a promising solution to support locomotion by providing assistive forces while preserving natural gait mechanics. However, most existing control strategies rely on predefined parameters and do not account for the user’s real-time physiological state. This work presents a hip-assistive soft exosuit integrated with a human-in-the-loop adaptive control framework based on wearable physiological sensing. Heart rate and respiratory activity, acquired through non-invasive commercial sensors, were combined into a fatigue-related index used to modulate level of assistance in real time, enabling support to be dynamically tailored to the user’s moment-to-moment physiological demand. The system was evaluated through treadmill experiments including level and uphill walking under multiple assistance conditions. A dedicated validation study also assessed the accuracy of the wearable respiratory estimation algorithms against a reference metabolic measurement system using a multilinear calibration model incorporating anthropometric variables. The validation results demonstrated good-to-excellent agreement across respiratory parameters, confirming the reliability of the wearable sensing approach for real-time physiological monitoring. Results demonstrated that adaptive assistance significantly reduced net metabolic cost and cardiovascular strain during uphill walking compared to unassisted and fixed-assistance conditions, while preserving gait kinematics and maintaining high mechanical transparency.