An Intelligent Control Framework for Robot-aided Resistance Training Using Hybrid System Modeling and Impedance Estimation

This study presents a novel therapy control method for robot-assisted resistance training using the hybrid system modeling technology and the estimated patient's bio-impedance changes. A new intelligent control framework based on hybrid system theory is developed, to automatically generate the desired resistive force and to make accommodating emergency behavior, when monitoring the changes of the impaired limb's muscle strength or the unpredictable safety-related occurrences during the execution of the training task. The impaired limb's muscle strength progress is online evaluated using its bio-damping and bio-stiffness estimation results. The proposed method is verified with a custom constructed therapeutic robot system featuring a Barrett WAM (TM) compliant manipulator. A typical inpatient stroke subject was recruited and enrolled in a ten-week resistance training program. Preliminary results show that the proposed therapeutic strategy can enhance the impaired limb's muscle strength and has practicability for robot-aided rehabilitation training.