Patient-Specific Imaginary Motor Movement Classification of Eeg Signals and Control of Robotic Arm

Abstract

This paper presents development of a robotic arm that can mimic the mobility of human arm with the recorded EEG (electroencephalogram) signals from the brain. This is a project designed for people who have lost the ability to control their arms. The robot arm will work with the signals that will be produced by the EEG instrument simultaneously when the user thinks that he or she will perform certain movements. These movements can be classified using Deep Multilayer Perceptron (Deep MLP) classifier designed in Python with Keras library. Deep MLP network is trained using Backpropagation algorithm for each patient by using own data. Experimental results demonstrate the proposed method can classify different imaginary movements with high accuracy.