Proxy-based Sliding Mode Control And Its Application In The Lower Limb Robots For Rehabilitation

In the past 60 years,rehabilitation robots have evolved from laboratory research equipment to many commercial products and have made great contributions to improving the efficiency of rehabilitation training and life quality of patients.Although great achievements have been made,there are still some challenges in this field,and safety is one of them.In order to guarantee both good tracking performance and intrinsic safety of the robots,position control for lower limb rehabilitation robot is studied in this paper.The proxy-based sliding mode control(PSMC)is a compliant control algorithm proposed by Kikuuwe,a Japanese scholar.It can be seen as a combination of traditional sliding mode control and PID control,so it also has the advantages of the both.In the case of large tracking error due to unexpected circumstances,PSMC exhibits the smooth and overdamped dynamic recovery effect of the ideal sliding mode control algorithm.While in the normal tracking state,it can show the characteristics of PID control which continuously adjusts the control signal according to tracking error.Therefore,PSMC can guarantee the tracking accuracy and system safety at the same time.However,there are still some theoretical deficiencies in PSMC.The stability analysis of PSMC is based on the assumption of local passivity of PID control,which is difficult to be satisfied in practical systems.Moreover,the assumption itself has not been strictly proved.Based on the idea of PSMC,two kinds of new proxy-based sliding mode control algorithms are proposed in this paper.First,a model-based PSMC(MPSMC)algorithm for a class of second-order nonlinear system is proposed,and the stability of MPSMC is analyzed based on Lyapunov theory and passivity theory.Based on a controller designing method for a class of under-actuated system,an under-actuated system control inspired PSMC(UPSMC)is proposed,and the stability of the system is analyzed based on Lyapunov theory.Large model uncertainty and external disturbance exist in the human-exoskeleton system.In order to improve the robustness of the system,an online parameter self-tuning algorithm is applied to the PID controller which is contained in the original PSMC,and an adaptive proxy sliding mode control(APSMC)is proposed.The above three algorithms are used in artificial pneumatic muscles(a commonly used actuator for rehabilitation robots),knee and ankle orthoses,respectively.The former is a class of strong nonlinear systems,and the latter two are systems with large model uncertainty and external disturbances.In this paper,the related experiments are carried out from the perspectives of position tracking accuracy,robustness and system safety,to verify the effectiveness of the proposed algorithms.