Control Strategy for a Lower-Limb Rehabilitation Exoskeleton Based on Dual-Impedance Control

Abstract

Fixed assistance cannot match changing lower-limb motor capability, while excessive intervention can suppress voluntary effort. This paper proposes assistance-as-needed control using position-velocity dual impedance. The framework defines a motion-state parameter from trajectory error and human-robot interaction force; in the present simulations. Safety limiting and admittance control convert normal corrective assistance and tangential velocity guidance into a reference trajectory. Across four virtual capability conditions, the prescribed parameter decreases from 0.75 to 0.25, desired velocity increases from 0.18 to 0.48 m/s, and relative velocity error remains below 1.67%. The model predicts reductions in normal and tangential assistance from 62.4 to 14.2 N and from 21.3 to 8.4 N, respectively. Under the highest-capability condition, virtual-model active mechanical work reaches 52.7 J. These results demonstrate downstream controller behavior under idealized simulation assumptions.