Sensor-Based Adaptive Robust Torque Control for Flexible Joints

Abstract

Achieving fast transient response and high steady-state tracking accuracy of joint torque for flexible joint systems with dynamic constraints, various parameter uncertainties, and uncertain nonlinearities is always challenging. To this end, a torque control scheme based on multi-torque sensors is proposed. After analyzing and transforming the joint dynamics, a lumped disturbance signal, including the reducer drive's torque loss, is constructed, which can be measured by the torque sensors. Then, a nonlinear disturbance observer (NDOB) and an adaptive robust torque control (ARTC) law are introduced to synthesize the control scheme, where NDOB estimates the motor's friction, and ARTC realizes the stable tracking of the joint torque. The performance of the proposed controller is theoretically ensured by Lyapunov analysis. Finally, a series of hardware experiments are carried out. The results suggest the effectiveness and superior performance of the proposed method for joint torque control by comparing it with other traditional methods.