Evaluation of the OpenBCI Neural Interface for Controlling a Quadrotor Simulation

Abstract

This thesis presents an initial analysis on the use of electroencephalography and electromyography to control the thrust settings of a quadrotor. The OpenBCI neural interface is used to sample muscle activity on a subject's face. Signal processing and event detection algorithms are implemented to identify eyewinks, and these wink events modify the thrust commands in a high fidelity, nonlinear quadrotor simulation. Currently only right and left wink events are detected; these can be mapped to two quadrotor commands such as fly up and down, roll right and left, pitch up and down, or yaw right and left. The ultimate goal of this project is to create a low-cost brain-machine interface system to fully control a real quadrotor using only bioelectrical signals such as electroencephalography and electromyography. A successful demonstration of the OpenBCI system may result in brain-machine interfaces that can be used in the development of low-cost prosthetic arms and legs.