MYO Armband sensors and Neural Network Algorithm for Controlling Hand Robot

Abstract

To control the robot hand can be used several methods; one of them is by using EMG sensor and time domain methods. In this study, Myo Arm Sensors combined with Neural Network algorithm are used. The Root Mean Square of the sensor signals is used to be learning by the system. The learning rate is 0.7 with two hidden layers. Each layer used three nodes. The results obtained that the system enabled to control robot real time a delay of around 1S. Moreover, the accuracy of the feedforward process in backpropagation Neural Network is 92.68%.