Implementation of Brain Computer Interface (BCI) as a Smart Wheelchair Motion Commands

Abstract

A wheelchair is a tool used to assist people with physical limitations in their legs. The most widely used are standard wheelchairs with a manual operating system by being pushed by hand. However, people with disabilities who have paralysis or suffer from neuromuscular and neurological conditions cannot use this wheelchair. Because of this, in this study focuses on implementing the Brain Computer Interface system to generate five commands to move a wheelchair. There are five important stages in the BCI system, that is signal acquisition, pre-processing, feature extraction, classification, and applications interface. Fast Fourier Transform (FFT) method used to extract brainwave features. The results of FFT are alpha (8-12Hz) and beta (12-30 Hz) waves in the frequency domain. For classifying brain waves into six classes as input commands to drive a DC motor used Support Vector Machine (SVM) method. Based on the test results, the average accuracy of the classification for the whole class reached 93,1%, the accuracy of class 0 (77,3%), class 1 (95,7%), class 2 (97,8%), class 3 (98,0%), and class 4 (97,5%).