Tuning the Magnitude and Timing of Exoskeleton Assistance using a Proportional-Joint-Moment Controller Improves Human Walking Performance

Wearable robots have demonstrated the ability to reinvent human mobility and reduce the biological effort of human movement. Proportional assistance strategy based on joint-moment-estimation has the potential to assist wearer in a wide range of human activities. The aim of this study is to systematically evaluate the physiological activities under different assistance magnitudes and timing. We measured soleus (SOL) electromyography, ankle angle, and biological ankle moment while participants walked on a treadmill using an ankle exoskeleton. Sixteen assistance conditions, combined with four torque levels (K = 0.1, 0.2, 0.3, and 0.4 relative to biological ankle moment) and four assistance timing (phase delay D = 0, 1%, 2%, and 3% of gait cycle), were applied to assist plantarflexion. In experiments with three healthy subjects, the assistance mode resulted in a maximum 31.1% reduction in SOL activity; a 13.71-degree increase in plantarflexion angle; and a 50.0% reduction in peak biological ankle joint moment compared to unassisted conditions. The experimental findings systematically characterize the dynamics of proportional assistance strategies, establishing theoretical foundations for optimizing exoskeleton control implementations.