A Hybrid Brain-Computer Interface for Smart Car Control

Abstract

Brain-computer interface (BCI) systems are often used to convert signals from brain activities into control commands through external devices. There are few studies on controlling a car by multimodality due to its difficulty in the current research. This paper proposes a hybrid BCI control system based on electroencephalography (EEG), electrooculography (EOG), and gyroscope signals to address this challenge. The user can control the start, stop, turn left, turn right, acceleration and deceleration of the smart car by this system. The user controls the start and stop by double blinking, acceleration and deceleration by concentrating and distracting, turning left and right by the head rotation. To evaluate the performance of this BCI system, we invited twelve subjects to conduct two online experiments to control the car on a runway to test the above functions. The experimental results showed that the hybrid BCI system achieved an average accuracy of 97.65%, an average information translate rate (ITR) of 43.50 bit/min, and an average false positive rate (FPR) of 0.70 event/min, thus demonstrating the effectiveness of our proposed system.