Development of an Upper Limb Master-Slave Robot for Bimanual Rehabilitation

Most activities in daily life involve the use of both hands. Post-stroke patients lose inter-limb coordination and require intensive rehabilitation therapy for a faster recovery. Robot based rehabilitation is preferred to traditional rehabilitation methods due to precise repeatability and the accurate measurements it provides. Bimanual rehabilitation robots can help the patient to keep the position symmetry. Most studies in rehabilitation robots have overlooked the necessity of accurate resistance training. Conventional rehabilitation methods cannot ensure position synchronization during rehabilitation. In order to realize resistance training with maximum efficiency, the resistance should be varied with a reference to a strength profile of a healthy person. This paper presents a master slave robot system designed for bimanual rehabilitation training. Disturbance OBservers (DOB) are implemented to make the system robust. Accurate force sensing is achieved using the Reaction Torque OBserver (RTOB). The paper also presents a method for seamless transfer between the resistance and assistance modes. The proposed methods were validated by experiments.