Tactile P300 Brian-Computer Interface Paradigm for Robot Arm Control

Abstract

This paper presents a tactile BCI for movement control of a three degree-of-freedom(DOF) robot arm. Six physical stimulus modules were placed on the left arm of the participant to stimulate event-related potentials with time-domain characteristics. Six different commands could be exported by classifying the Electroencephalogram(EEG) signal, which could control the robot arm movement in the corresponding position. To test the practicality of our BCI system, 5 subjects took part in the experiment with their appropriate stimulation time, electrode channel and corresponding classifier parameters. In the online experiment, the average of classification accuracy was 84.7%, with an average information transfer rate at 12.2 bits/min. In the robot arm control experiment, all subjects could complete a reach-to-grasp task. The results show that the tactile BCI paradigm is feasible and practical for some multi-DOF control applications.