Torque-Tracking Control Combining Polynomial Feedforward With Sliding Mode Feedback for Magnetorheological Actuators

The Magnetorheological (MR) fluid-based actuators have received much attention for the advantages of high torque-to-inertia ratios, low noises, and low energy consumption. Despite these unique characteristics, the applications of MR actuators are trapped by precise control due to the uncertainties and disturbance. This article proposes a new double-loop control scheme by combining the model-based feedforward with the signal-based feedback for the fast and accurate torque-tracking control of MR actuators. A polynomial model for feedforward control is built based on a rotary MR actuator to establish the relations between the torque and current. Considering the system uncertainties and disturbance, a discrete sliding mode compensation control is used to reduce torque-tracking errors according to the output torque feedback. A piecewise sliding mode method with variable parameters is designed to reduce the chattering of the sliding mode control. A test bench is set up for parameters identification of the polynomial model. After parameters identification, the proposed double-loop control method is verified with the experiment. The results show that the double-loop control achieves good torque-tracking performance with a maximum root mean square error lower than 0.54 N·m. Moreover, this torque-tracking effectiveness improves greatly when the piecewise sliding mode control with variable parameters is introduced for the chattering control. Compared with the feedforward-based PID control, the proposed sliding-mode-based double-loop control method has higher accuracy.