Soft Robotic Glove with Motion Controlling System Based on BP Neural Network

In order to facilitate the rehabilitation of patients with hand dyskinesia, a soft mechanical glove with the master-slave motion controlling system is proposed in this paper. Since the relationship between the input collected from movement of master hand and the outputthe parameters needed to control slave glove is complex, the combination of the motion controlling system and BP neural network is applied in this paper to ensure the accuracy of simulating motion for slave glove. Firstly, the hardware and software design of the glove is demonstrated. And then, the training data of the fingertip-palm-distance and the corresponding rotating angles of steering to simulate the movement of master hand is calculated, based on which the BP neural network is established. Finally, the motion control error is proven within 5% by experimental validation. The BP neural network based controlling system demonstrated in this paper can also be applied in other similar problems in engineering practice.