Upper limb rehabilitation robot integrated with motion intention recognition and virtual reality environment

Robotic-assisted therapy is of benefit to the recovery of upper limb motor function for the patients survived stoke. Whereas, there are few emphases on the patients' motion intention during the rehabilitation process. The goal of this study is to combine the control strategies based on patients' motion intention with an upper limb rehabilitation robot to improve the recovery for the patients. In this paper, we propose an integrated robot-assist rehabilitation system, in which a 3 degree-of-freedom (DOF) exoskeletal rehabilitation robot, an EMG-based intention recognition module and a VR game environment are seamlessly combined. According toharacteristics of EMG signals, the wavelet package analysis approach is applied to extract the features of EMG. The node energy is used to construct the feature vector instead of the original coefficients of wavelet package decomposition to resolve the time-invariance problem. Then feature projection results in the singularity problem of with-in scatter matrix during the feature dimension reduction. To overcome the disadvantage of the with-in scatter matrix, this paper uses a recursive algorithm which is proposed in our previous work. The reduced feature vector is recognized by a neural network classifier and the output of the classifier is used for the control inputs. Preliminary experiments are also performed to implement the control of the rehabilitation robotic system by using the proposed EMG reorganization method, together with a dart game realized in the virtual reality environment. Experimental results show that the performance of motion intention recognition is satisfactory and the entire integrated system is feasible.