Time-delay Positive Feedback Control for Nonlinear Time-delay Systems with Neural Network Compensation
基于神经网络补偿的非线性时滞系统时滞正反馈控制

A new adaptive time-delay positive feedback con- troller(ATPFC)is presented for a class of nonlinear time-delay systems.The proposed control scheme consists of a neural networks-based identification and a time-delay positive feedback controller.Two high-order neural networks(HONN)incorpo- rated with a special dynamic identification model are employed to identify the nonlinear system.Based on the identified model, local linearization compensation is used to deal with the un- known nonlinearity of the system.A time-delay-free inverse model of the linearized system and a desired reference model are utilized to constitute the feedback controller,which can lead the system output to track the trajectory of a reference model. Rigorous stability analysis for both the identification and the tracking error of the closed-loop control system is provided by means of Lyapunov stability criterion.Simulation results are in- cluded to demonstrate the effectiveness of the proposed scheme.