Learning-Based Adaptive Impedance Control for a Human-Powered Augmentation Lower Exoskeleton

A learning-based adaptive impedance control algorithm for a human-powered augmentation lower exoskeleton (HUALEX) is presented. The HUALEX system architecture is introduced first, which is divided into three parts including the mechanical subsystems, the sensing subsystem and the control subsystem. By using impedance control method, the inverse dynamics model of HUALEX is established and the control effect of impedance parameters is studied. And then, a reinforcement learning-based adaptive impedance control algorithm, including the reinforcement learning, PI² (policy improvement with path integrals) learning algorithm and adaptive impedance control, is proposed. The effectiveness of the algorithm is verified simulation experiment.