Investigations of a novel lower-limb exoskeleton control strategy

Abstract

A novel controller was developed with a single positional input to determine the torques for each joint of a hip-knee-ankle leg of an exoskeleton during stance, and a regenerative minimum jerk based swing control policy. Simulations were performed in MATLAB, and Simulink environments. These simulations involved using a dynamic model to evaluate if the novel controller could produce natural gait patterns. Additionally, simulations evaluating regenerative minimum jerk methods were performed in Simulink to investigate swing control applicability. The stance control results demonstrated that proper progression through stance sub-phase is generated along with predictable torque outputs, however, the lack of synchronization between legs leads to uncoordinated gait. Swing control simulations demonstrated successful implementation of a regenerative minimum jerk control policy on a three degree-of-freedom leg. Moving forward, a high-level controller could be used to finalize the control framework by ensuring synchronization between the independent legs.