Adaptive Neural Network Command Filtering Technique for Resolving a Class of Markov Jump Systems with Dead-Zone Inputs

In this paper, the method of command filtering based on adaptive neural network is used to solve the problem of Markovian jumping nonlinear system with dead-zone input. Meanwhile, the command filtering method is introduced into this system to reduce the computational burden. Moreover, due to the existence of unknown smooth nonlinear switching terms, we use the common neural network technology for fitting. And the dead zone problem of actuator will be solved by decomposition of its expression form. Finally, a novel neural network tracking control structure is constructed and the validity of the controller is verified specifically for the second order nonlinear switching system.