Implementation of Artificial Neural Network in Electric Motor Control using Brain-Computer Interface

Abstract

Brain-Computer Interface (BCI) is a technology that integrates analog brain signals to digital computer-based systems for the purpose of analysis, manipulation, and control. With modern machine learning algorithms, Artificial Neural Network (ANN) was used to develop the intelligence in processing brain signals for electric motor control. In this study, supervised and unsupervised learning methods were explored to train the ANN. Guided and unguided thought-Task methods were used in manipulating a 5 and an 8 switching control variations of an electric motor. A brain wave filter was designed, and the different brain bands were explored. Significant signal features were obtained and were varied in terms of qualification. Simulation run in an offline and real-Time mode. Results show high control accuracies in using the Gamma band with a supervised learning method. Guided thoughts with 5 switching controls, and 4 features gave better results. Control accuracies varies between off-line and real-Time implementations.