Effects of User Intent Changes on Onboard Sensor Measurements During Exoskeleton-Assisted Walking

Adoption of exoskeletons for rehabilitation and locomotion outside the physical therapy clinic has been limited by relatively primitive methods for identifying and incorporating user intentions. Gait intention resolution varies from high-level goals (increase speed) to mid-level actions (increase stride length) to low-level joint behaviors (increase hip flexion). Onboard sensors only indirectly sense the human via the exoskeleton interface, but offer measurement consistency advantages over more direct methods. In this study, exoskeleton users, both able-bodied and having spinal cord injury, performed goal-level changes in gait speed to characterize joint- and action-level responses. Trials were completed in both trajectory-free and trajectory-based control modes, with both crutches and a walker. Results confirm statistically significant differences between the pre- and post-speed change joint-level measures of position and motor current. Coordination of joint-level changes resulted in significant differences in action-level measures (stride length and stride time). Users realized goal-level speed changes of as much as 0.30 and 0.19m/s in the trajectory-free and trajectory-based modes, respectively. Findings suggest that intent detection is possible for able-bodied and non-able-bodied users with onboard sensors alone. The intent signals depend on exoskeleton control settings, user ability, and gait phase, but do not differ with use of crutches or a walker. The characterization of these intent-related signals via onboard sensors enables more detailed intent recognition without the need for external sensors, which could benefit any control strategy that explicitly incorporates user intent.