Adaptive Control of a Hysteretic Magnetorheological Robot Actuator

In this paper, a new adaptive control scheme is proposed to compensate for the magnetic hysteresis in magnetorheological (MR) fluid-based actuators. MR actuators offer high torque-to-mass and torque-to-inertia ratios. Input and output shafts are mechanically decoupled in MR actuators, providing lower inertia compared to the same power geared motors. Additionally, geared motors typically add significant noises on the torque/force measurements. Noises on the torques/forces can be attenuated in MR actuators due to fluidic connections between the input and output shafts, facilitating high fidelity torque/force control. Despite these unique characteristics, magnetic circuits within MR actuators create hysteresis between the input current and output torque that negatively affects the quality of torque/force control and the repeatability of MR actuators. A hysteresis compensation scheme is essential to gain repeatable and high-quality actuation. To this end, we propose an adaptive control method that estimates both hysteresis and uncertain parameters of the magnetic circuit and cancels nonlinearities based on feedback linearization technique. A set of experiments is performed to validate the effectiveness of the proposed method, and the results are compared to a proportional-integral-derivative controller.