Asynchronous Motor Imagery Brain-Computer Interface for Simulated Drone Control

Abstract

Brain-computer interfaces allow direct control over devices without any physical action by the user. Motor imagery-based brain-computer interfaces analyze spatial patterns from brain signals elicited when the user imagines execution of a specific behavior. One of the ways to obtain such brain signals is with electroencephalography, which measures signals over the scalp. In this paper, we analyzed the brain patterns from when the users performed different motor imagery tasks and applied them to navigate a simulated drone. The drone was controlled asynchronously, with the user's intentions continuously analyzed throughout the entire drone control period. By navigating the drone in two different scenarios using either 4 or 6 control commands and by comparing control performance when controlling the drone with either a BCI or a keyboard, we have shown the feasibility of motor imagery for asynchronous control of drones for both two- and three-dimensional device control.