Extracting and discriminating selective brain signals in non-invasive manner and using them for controlling a device: A cost-efficient approach to brain computer interface (BCI)

Abstract

The interface through which a human brain establishes links with external devices is generally called Brain Computer Interface. Although there are some significant amounts of ongoing researches on how an overall efficient BCI can be developed are going on, making a cost-efficient approach while dealing with limitless brain patterns is found to be more challenging. In this work, feasibility of a cheaper but appropriate way of extracting and discriminating of several non-invasive EEG signals and using those for controlling devices such as a wheel chair has been proved. To assist the argument of this project, numerous experimental data has been processed to produce several signals, such as, right turn, moving forward, stop etc. for the wheel chair. In the experiment the above mentioned three signals were well distinguished from each other. A microcontroller has been used for processing the signals collected from the brain and hence sending to the wheel chair controlling motors. Despite the challenges of dealing with very low but noise sensitive brain signals, their limitless patterns, and limited scope of necessary circuitries, this work has opened up the scope of feasibility of BCI technology in practical life with a simpler and easier approach.