Control of a Quadcopter with Hybrid Brain-Computer Interface

Abstract

Brain-computer interface (BCI) devices are designed to bypass neuromuscular pathways enabling an individual to operate an external device using neural activity instead of motor activity. The use of BCIs in robotics control has important applications of assisting patients with severe motor disabilities, such as exploration, home assistance, and control of an electric wheelchair. This paper presents a development of a practical implementation of a BCI by utilizing low-cost technologies and asynchronous signal processing techniques for EEG signal acquisition and processing to navigate a quadcopter. A wireless EEG headset was used to acquire the occipital alpha and the frontal muscular artifacts from a user along with the gyroscope signals from the headset. These signals were transmitted wirelessly to a PC for further processing. Signal-processing algorithms were developed and implemented in the PC to extract the patterns from the acquired signals. These patterns, representing the intentions of the user, were then used to control a quadcopter. The design was successfully tested in a virtual environment and a physical device. The results from this preliminary pilot study proved that the design could be used with high accuracy to control a robotic device.