Neural Network Based Adaptive Memoryless Observer Design for Nonlinear Time-Delay Systems

• 作 者: Zhi-Jun Fu Wen-Fang Xie
• 机 构: College of Mechanical Engineering,Zhejiang University of Technology,hangzhou,310014,China
• 出 处: 第26届中国过程控制会议
• 发表日期: 2015年10期
• 关键词: Nonlinear observer time-delay neural network memoryless observer

　　In this paper,a neural networks based adaptive memoryless observer for nonlinear time-delay systems is proposed.The proposed observer is based on nonlinear-in-parameter neural networks which makes it applicable to many systems with arbitrary degrees of nonlinearity.The Lyapunov function technique is used to develop a novel stable updating law for the neural networks model and to prove that the state error,output estimation error and the neural network weights errors are all uniformly ultimately bounded(UUB)around the zero point.Moreover,the precise system model,the strictly positive real(SPR)assumption,the Lipschitz,linear-in-parameter or norm-bounded assumptions of nonlinear functions are not needed.Robust property of the proposed observer is proved by passivity method.Compared with other nonlinear observers,the proposed observer with the newly designed simple updating law demonstrates faster convergence speed.Two numerical examples are also given to demonstrate the effectiveness of the theoretical results.