Abstract:

：A certain limitation still exists in the current robot-aided resistance training method, in which the therapeutic controller is mainly designed based on the impaired limb’s active force and electromyogram signal which is vulnerable to noise impact, and further it does not incorporate continuous variable motion control and discrete event decision control into a unified framework. To solve the aforementioned problem, a therapy control method based on impedance identification and hybrid modeling is proposed. Firstly, the impaired limb’s time-variant bio-impedance is online identified by using the impaired limb’s active force and actual position. Secondly, the robot-aided resistance training system model is built using hybrid control theory, in which the discrete events for continuous dynamic system is defined according to the identified impaired limb’s bio-impedance changes, moreover, the discrete control states and the discrete event supervisory controller based on hybrid automation are also given for discrete dynamic system. Finally, the experimental system platform is constructed by using the WAM^TM compliant manipulator of Barrett Inc., and the experimental results verify the effectiveness and practicability for the applications to robot-aided resistance training using impedance identification and hybrid control theory.

Key words: hybrid control, impedance identification, resistance training, rehabilitation robot