Robust Position Control of Series Elastic Actuator with Sliding Mode Like and Disturbance Observer

Series elastic actuators (SEAs) are powerful devices in the area of human-robot collaboration. However, they still suffer from difficult position control issues, mainly because of the increased model complexity inherent in the actuation and dynamics. Therefore, an efficient approach is proposed to solve this problem in this paper, in which a disturbance observer (DOB) and a nonlinear robust sliding mode like (SML) controller are synthesized. The DOB is designed to estimate not only the modeling uncertainties like parameters perturbation but also unmodeled dynamics; the nonlinear robust controller is designed to attenuate the estimation error and stabilize the closed loop system. The boundedness analysis is presented and the global convergence is analytically proven. The performance of the proposed approach is evaluated on a self-built SEA platform and experiment results illustrate the effectiveness of the proposed scheme.