Real-Time BCI Scheme to Assisting Disabled Persons Using Motor Imagery EEG Signal Classification

Abstract

Electroencephalography (EEG) signal-controlled Brain-computer interface (BCI) schemes have created hope for physically impaired people to lead a stress-free life. It is quite challenging to preprocess EEG signals and make them eligible for use in neuro-robotics applications as there exist various categories of artifacts in the raw EEG signal. As physically disabled people need to perform real-time actions, this study proposes a real-time BCI scheme that is usable and efficient for neuro-robotics applications in their rehabilitations. The ultimate goal is to classify hand, foot, and tongue motions as four-class motor imagery (MI) task-related impulses using the more efficient classification technique in this study. The proposed approach achieves state-of-the-art levels of accuracy and a kappa score of 77.41% and 0.70, respectively, on the benchmark dataset taken. Classifiers such as Skl-ANN, SVM, LDA, and others have been evaluated for the experimental subjects. It is believed that the higher classification accuracy and lower processing load of the proposed BCI system will make it acceptable for usage in real-world settings to facilitate the rehabilitation and reintegration of physically impaired persons.