Design and Optimization of a Real-Time Asynchronous BCI Control Strategy for Robotic Manipulator Assistance

Abstract

This paper presents the design and evaluation of a robotic system controlled by motor imagery based Brain-Computer Interface (BCI), which allows users to perform reach-grasp-release activities in 3D space. The sequential axis control methodology is used in the research to address a low information-transfer rate problem, where a participant has to provide several low-level commands to the robot, such as position on x, y, and z-axes, and close/release gripper actions. With the advantage of the motor imagery paradigm, the system is asynchronous, and the operator performs self-paced control. The BCI system is based on sensory-motor rhythms elicited by electroencephalographic (EEG) signals in the frequency domain to classify mental intent in real-time by a soft voting ensemble classifier model. The experimental results have shown the feasibility of controlling a six-degree-of-freedom robot manipulator with the classification performance in the range of 64 - 86% tested on seven healthy individuals.