Adaptive neural network decentralized stabilization for a class of uncertain nonlinear time-delay large-scale systems with expanding construction

Abstract

An adaptive neural network decentralized stabilization approach is presented for a class of uncertain nonlinear time-delay large scale interconnected systems with expanding construction in this paper. So-called expending construction of large scale interconnected systems means that there are some new subsystems to be added in the original large scale system during operation. The control problem is to design the decentralized control law of the newly added subsystem on the basis of not changing the decentralized control laws of the original system, so that the new subsystems and the expanded nonlinear time-delay large scale system are all stable. In this paper, the decentralized control law and the parameter adaptive law of the new nonlinear subsystem are designed by using backstepping technique. And RBF Neural Networks are used to approximate both the nonlinear functions of the subsystems and the delay-dependent nonlinear functions in the interconnections. An adaptive NN output feedback controller is constructed. By constructing a proper Lyapunov functional, it is proved that all signals in the expanded closed-loop nonlinear large scale system is bounded. An illustrative example shows the feasibility of the presented approach.