Brain-Computer Interface based on Neural Network with Dynamically Evolved for Hand Movement Classification

Abstract

Translating brain commands into movements on the prosthetic robot is not an easy task. It is needed a control system for the prosthetic robot based on human body signals to predict the desired movement so that the robot is part of the body. This assistive device is used to help people with disabilities perform functional movements such as gripping with motor activities performed on all five fingers. This paper proposed a hand movement recognition system based on electroencephalogram (EEG) using the Neural Network with Dynamically Evolved Capacity (NADINE). The data generated from the model test shows almost the same value as NADINE, with a maximum accuracy of 98% and an average prediction time of 14 milliseconds. These results further strengthen that the NADINE model can be used in real-time.