Nonlinear System Of Intelligent Sliding Mode Control Method

The intelligent sliding mode variable structure control (ISMVSC) has been deeply studied as a nonlinear control scheme. In this paper the current ISMVSC theory and methods have been investigated, and several sliding mode control schemes integrated the genetic algorithm, fuzzy control, and the neural network are proposed and implemented for a class of nonlinear system.Firstly, sliding mode controller parameters are regulated by genetic algorithm in real-time to construct the optimal switching function and exponential approach law factors. And the hyperbolic tangent function will instead of the saturation function of traditional method to suppress the chattering. Therefore the control system has good dynamic property, faster reaching ability to the switching surface, and stronger robustness. The simulation results in the inverted pendulum system proved the effective of this scheme.Next, a terminal sliding mode control method based on genetic algorithm is studied. The genetic algorithm is used to select the optimal parameters of the fast terminal sliding mode controller, which can insure that the system state will reach rapidly and maintain the terminal sliding surface in the finite time. And the scheme can avoid singularity case in the control progress, keep good robustness and weaken the chattering simultaneously. The simulation experiment with this method on a robot manipulator is implemented and good system performance is achieved.Moreover, using the fuzzy neural network (FNN), a new intelligent terminal sliding mode controller is designed. The FNN is adopted to regulate the switching gain, at same time, the system modeling error and the disturbance upper bound is evaluated automatically using an integral method to realize the automatic tracking and chattering attenuation. The scheme is applied to a two-degree-of-freedom robot system, and simulation results show that the system has good dynamic performance and stronger robustness.