Adaptive fuzzy sliding mode controller design for a new hand rehabilitation robot

Abstract

Hand rehabilitation is one of the most important rehabilitation procedures. Due to the repetitive nature of rehabilitation training, a full robotic system could help the physiotherapists to gain time for creating new training schemes for a larger number of patients. Such a system can be based on live or recorded data and consists of the operator-device, patient-device, and control mechanism. This paper focuses on the design of the patient-device and its control-system in a decoupled training scenario. It presents a robot for hand rehabilitation training fingers and wrist independently based on only two actuators. These two actuators are configurable to allow consecutive training on the wrist and all joints of the fingers. To overcome uncertainties and disturbances, a sliding mode controller has been designed and an adaptive fuzzy sliding mode controller is used to reduce the chattering effects and compensate the varying forces of the patients. The experimental results show an approximate 80% improvement in tracking the desired trajectory by the adaptation.