Error-Induced Brain-Actuated Motor Command Generation for Speed-Profile Setting in a 2-Loop Position Control of a Robot Arm

Traditional BCI-based position control schemes utilize acquired motor-imagery signals of the subject for motor actuation and error-related potential signals to detect positional overshoots. The positional overshoot results in steady-state positional error. Current BCI research attempts to reduce steady-state errors at the cost of high cognitive load of the subjects. The motivation here is to reduce the steady-state positional error without enhancing cognitive load of the subjects. A 2-loop position control scheme is proposed here with an outer loop for position control of the motor and the inner loop for speed-setting of a robotic limb. Three distinct models of brain-actuated speed-setting with possible control policies and stability analysis are presented. The analysis reveals that Takagi–Sugeno-type fuzzy control yields the best results. Additionally, cognitive load is also reduced significantly as the user needs to generate error-related potential once only to identify the position of the targeted object he needs to translate.