Research On Chattering Suppression In Sliding Mode Variable Structure Controller For Industrial Robot Manipulator

Industrial manipulator is a typical nonlinear time-varying system, which is strong-coup led and with much uncertainties. Because of strong robustness against disturbances and variation, sliding mode variable structure control is an efficient and common approach in robot manipulator control. However, Its application and development are limited by the chattering phenomenon.In order to alleviate chattering, merging sliding mode variable structure control with artificial neural networks and fuzzy system are studied in this paper.First, the development of the manipulator and manipulator control methods are reviewed. Then the basic principle of sliding mode variable control is introduced and designing process is discussed in detail.And a two-power fast terminal sliding mode control of manipulator is proposed. It is robustness against parameter variation and can be physical implemented easily. And, compared with the linear sliding mode or terminal sliding mode, it can make the system states converge to zero in faster speed.The designing and learning method of artificial neural network are studied to avoid chattering phenomenon in sliding mode control system. And sliding mode variable structure control based on neural networks compensation for manipulators is proposed. It uses neural networks to online identify the upper bound of uncertainties. And the neural networks are trained by gradient descent method. The stability of controller is proved by Lyapunov’s theorem of stability. Simulation results verify the validity of the control scheme. Even mo re,the chattering is reduced and system robustness becomes stronger.Designing method of fuzzy system is studied, and a sliding mode variable structure approach with fuzzy switch gain for manipulator is proposed. In this method, first the fuzzy relation between the sliding mode function value and switch gain in control input, then appropriate fuzzy systems are designed to adjust the switch gain online. It reduces the chattering phenomenon and improves the system robustness. The parameter update law is deduced by using lyapunov direct method, which ensure that the control system is stable.