Neural Network Control For Two Classess Of Switched Nonlinear Systems

In recent years,switched stochastic nonlinear systems have obtained great attention in the area of control,which have important scientific theoretical and practical significance.However,there are many areas that have not been explored yet.In this paper,the observer and controller are territory from the perspective of non-lower triangle and input saturation problems:1.For a class of the un-lower triangle switched systems,the approximation capability of the neural network and the dynamic surface control combined Backstepping technique,Through the adoption of all subsystems of the subsystems of the common coordinate transformation instead of coordinate transformation.Using the improved average dwell time,we show that the overall closed loop system is bounded stable and the output of the switching system converges to a small neighborhood near the origin.2.A class of input saturated switching systems is studied,the design difficulty caused by input saturation is solved,Using the variable separation techniques overcomes the unstrict feedback structure design problem,and through the backstepping technique to construct the common lyapunov function.Further,we can obtained the closed-loop systems is stable of uniformly bounded.Finally,the correctness of the two classes of control schemes is confirmed by a numerical simulation of the designed switching system.