DESIGN AND DEVELOPMENT OF BIOSIGNAL CONTROLLED HAND EXOSKELETON FOR ASSISTIVE PURPOSES

Abstract

Exoskeletons are widely used for both rehabilitative and assistive purposes. Controlling the exoskeleton remains significant to perform the user-defined functions. This work presents the design of a low cost, biosignal-controlled hand exoskeleton exclusively meant for assistive purposes. A biosignal acquisition unit is designed and developed to acquire the electromyography (EMG) signals from Biceps, Extensor Digitorum, and Flexor Digitorum Muscles. Two different hand movement protocols are used to obtain the EMG from 21 healthy individuals. From the acquired signal, 53 features including higher order statistical features are extracted and fed to different classifiers for the classification of three different hand movements. For both the protocols used, the decision tree classification method shows a higher accuracy of 90.47% and 95.23%, respectively, compared to other classifiers. The exoskeleton is designed, 3D printed, and assembled with linear actuators. The classifier output controls the device to perform hand movements. Different hand movements specified in the protocol are executed by the exoskeleton fitted in a normally functioning hand with the help of actuators controlled by the biosignals acquired from the other hand. The inclusion of brain signals is expected to provide greater accuracy in extracting the control signal.