Surface Electromyography-Controlled Mechanical Hand: A Design Approach

Abstract

As the human hand is a biomechanical system, its function and motion can be replicated artificially as a mechanical system. Mechanical hands can be used to perform dexterous tasks performed by humans in an environment unsuitable for humans, or as prostheses, etc. In doing so, a user interface like a man–machine interface is required to control the development of motion of the mechanical hand. One such method uses surface electromyography (sEMG). The sEMG signals can be modeled to predict intended movement. The work presented here is focused on developing three blocks namely the mechanical design, intent analysis, and movement control block. To this effort, a 16 degree of freedom (DOF) mechanical hand is designed. The intent analysis is done using regression neural network. For the movement control block, an embedded system is developed to detect sEMG signals and generate control signals for the actuators of mechanical hand. Overall in this work, all three blocks are integrated to develop a design of a sEMG-controlled mechanical hand.