Model-Free Non-Singular Fast Terminal Sliding Mode Controller Using Extended Sliding Mode Observer for Series Elastic Actuator-Based Robotic Arm

This paper proposes an adaptive admittance controller and a model-free non-singular fast terminal sliding mode controller based on an extended sliding mode observer for hybrid force and position control of a robotic arm driven by series elastic actuators (SEA) with external disturbances and model uncertainty. According to the ultra-local model method, simplifying the robotic arm driven by SEA dynamic model and all unknown terms which include the uncertainty of the model and unknown disturbance are integrated into one term called lumped total disturbance. In order to estimate the integrated unknown disturbance terms, an extended sliding mode observer (ESMO) is constructed. For better position control performance, a model-free non-singular fast terminal sliding mode controller was constructed by combining ultra-local model with non-singular fast terminal sliding mode control. In the outer loop, in order to accurately control the contact torque generated by interaction with the external environment in the hybrid force and position control of the robotic arm with changes in external environmental stiffness, an adaptive admittance controller is designed. To validate the effectiveness of the control strategy and the optimization effect on control performance, co-simulations of the SEA driven robotic arm are realized by using SolidWorks and MATLAB. Compare the co-simulation results with TDE-iPD to show the high accuracy and fast convergence characteristics of this control method.