Adaptive neural network decentralized control for nonlinear interconnected systems with time-varying constraints

This paper focuses on a decentralized adaptive scheme for a class of nonlinear time-varying interconnected systems based on neural networks for output tracking. By introducing constraint estimation method and two smooth functions, the barrier of unknown interaction in the system is avoided. By means of incorporating the back-stepping technique and the capability of neural networks, we aim to approximate the unknown nonlinear parts and establish a novel decentralized scheme with prescribed performance. Furthermore, according to Lyapunov stability theorem, it can be inferred that all variables of the controlled system are bounded, while the expected tracking converges to a compact set with a small error range. In addition, the simulation results verify that the proposed scheme can obtain a rapid learning effect.