EMG control of a 3D printed myo electric prosthetic hand

Abstract

Abstract Recent technological advances have enabled the prosthetic developers to derive an ideal replacement for the human arm in near future. This research presents design of a 3D Printed Myo Electric Prosthetic Hand that grasp symmetrical objects through Electromyographic signals and is intended for people suffering from Trans-radial Amputation. In the current study, a set of data about three motions that include hand open, close and rest position was been acquired from forearm muscles of a human arm. The computer aided model of the prosthetic hand was developed in SolidWorks® which was later 3D printed using Poly Lactic Acid (PLA).The proposed design was based closely on tendon actuation, related to human hand functionality. Signal acquisition and processing has been done using Myo Armband. Different features were selected such that they can be passed through classification process in order to control the hand motions. Using the Bluetooth transmitter, the filtered data was sent and saved in Arduino Uno® controller. Later, the Support Vector Machine (SVM) was been evaluated as a classifier. The classification accuracy obtained from SVM was 96.7%. The results were found significant (p<0.01) for twelve able bodied subjects. EMG based grasp control was implemented with successful testing of designed prosthetic hand for three different motions. The practical effectiveness of the Myoelectric Hand was demonstrated by grasping household objects such battery charger, wallet and water bottle.