Hybrid Impedance-Sliding Mode Switching Control of the Indego Explorer Lower-Limb Exoskeleton in Able-Bodied Walking

Curt A. Laubscher,Anthony Goo,Ryan J. Farris,Jerzy T. Sawicki
DOI: https://doi.org/10.1007/s10846-022-01583-7
2022-04-01
Abstract:This paper proposes a novel hybrid controller for promoting safe human-robot interaction. The hybrid controller modifies a model-based impedance controller such that it uses impedance control but switches to sliding mode control under non-nominal conditions. Each control law is formulated with an inner-loop controller for feedback linearization and an outer-loop feedback controller for trajectory tracking. The outer-loop feedback torque is theoretically proven to have a smaller magnitude in hybrid control than in impedance control under an assumed condition, suggesting it may be the safer approach. To validate the mathematical assumption and purpose of the controller, a walking experiment is conducted where a healthy able-bodied subject using a lower-limb exoskeleton is randomly subjected to either hybrid or impedance control. Perturbations are induced through sudden changes in treadmill speed, resulting in operation outside nominal conditions for 15.9% of the experiment. The assumption made in the theory holds true for the majority of the experiment, failing only 14.3% of the time. The main results show a statistically significant reduction in average feedback torque magnitudes by 7.9%. This is accomplished without drastically affecting gait, with joint angle root-mean-square differences being 0.36° for the hip and 0.64° for the knee. This demonstrates how the hybrid controller can achieve similar gait patterns with lower feedback torque magnitudes, suggesting it is a promising alternative to impedance control.
robotics,computer science, artificial intelligence
What problem does this paper attempt to address?