A human-robot interaction based coordination control framework for walking assist

Abstract

In order to improve flexibility and multi-joints autonomy in walking assist, a novel notion of human-robot interaction (HRI)-based coordination control method is explored by combing a HRI technology based on central pattern generators (CPGs) network with an impedance structured control. Firstly, a CPG unit with its outer-interaction mechanism is connected to each left and right hip joint to guarantee the independent HRI occurring at each side. Secondly, mutual inner inhibitions are connected to both left and right CPGs to maintain an anti-phase relationship of the left and right hip joints, and thus to guarantee stable walking assist in the complicated HRI environment. Thirdly, a framework of impedance structured control is designed for knee joints, where a high and low impedance control law is designed for stance and swing phase respectively. Finally, a hip-knee linkage mechanism is implemented to establish relevant natural hip and knee joint movement of one unilateral leg in walking. Computer simulations are carried out to demonstrate the effectiveness of this proposed control framework.