Variable Stiffness Control of a Soft Rehabilitation Robot with Sliding Mode Method

In most applications that involve human-robot interactions, compliant actuators with variable stiffness can be used to improve fast and safe motion for robots. The bio-inspired agonist-antagonistic setup of variable stiffness actuator (VSA) systems enable online compliance adaptation through co-contraction of actuator pairs. However, the promise of these platforms is limited by the difficulty of control. In this paper, we develop a sliding mode variable stiffness controller for a soft rehabilitation robot actuated by an antagonistic pair of McKibben muscles. To demonstrate the stability of this controller, the robotic control system is analyzed using the Lyapunov method. Finally, several simulations are conducted on the robot, and the results show that the proposed controller has good tracking performance and robustness.