Recurrent Neuroadaptive Output Feedback Control of Motor Servo Systems With Disturbance Compensation

ABSTRACT A recurrent neuroadaptive output feedback controller which combines recurrent neural network (RNN) and extended state observer (ESO) will be proposed for motor servo systems with model uncertainties and signal noises. Notably, RNN can accurately approximate endogenous uncertainties. However, under strong external disturbances, its approximation performance may deteriorate and even instability may happen. Therefore, the controller in this paper will be proposed to handle these control issues. Simulation and experimental results demonstrate the effectiveness of the proposed controller.