4 DOF Exoskeleton Robotic Arm System for Rehabilitation and Training

This paper presents a rehabilitation and training system with 4 DOF exoskeleton robotic arm. This proposed system can record a posture of physiotherapist and playback that posture to the patients. For the posture playback, the exoskeleton arm's motion was controlled with the recorded gesture and adjusted the level of an assistive motion. The GRNN method was used for predicting the static gravity compensation of each joint with accuracy of 94.66%, 97.63%, 87.02%, and 97.32%, respectively. Hence, the exact system modelling was not required in this system. The force controller with admittance control method was applied to control this exoskeleton robotic arm. The results of the usability test showed that the proposed system had an ability to enhance the muscle's strength and indicated that the purposed exoskeleton arm could be applied to the rehabilitation or training task.