非线性零阶接近有界多模型神经网络自适应控制器  

Nonlinear adaptive controller using multiple models and neural networks based on zero order proximity boundedness

作　　者：黄淼[1] 王昕[2] 王振雷[1] 

机构地区：[1]华东理工大学化工过程先进控制和优化技术教育部重点实验室,上海200237 [2]上海交通大学电工与电子技术中心,上海200240

出　　处：《控制与决策》2013年第9期1315-1321,共7页Control and Decision

基　　金：国家自然科学基金优秀青年基金项目(61222303);上海市科技攻关项目(12dz1125100);上海市重点学科建设项目(B504);流程工业综合自动化国家重点实验室开放课题基金项目

摘　　要：针对一类单变量非线性离散时间系统,提出一种零阶接近有界的多模型神经网络自适应控制器.该控制器包含一个非线性鲁棒自适应控制器和一个非线性神经网络自适应控制器.当系统非线性项放宽到零阶接近有界时,这两个控制器分别用于保证系统的稳定性和提高系统的性能,系统的控制输入由切换机构在两个控制器之间进行切换产生.最后给出了稳定性和收敛性证明,并通过仿真实验验证了该控制器的有效性.A nonlinear adaptive controller using multiple models and neural network based on zero order proximity boundedness for a class of single variable nonlinear discrete-time systems is proposed.The controller is consist of a nonlinear robust adaptive controller and a nonlinear neural networks adaptive controller.These controllers can respectively guarantee the stability of the system and improve the performance of the system when the restriction of nonlinear term is relaxed to the zero order proximity boundedness.The control input can be generated by the action of the switching mechanism between the two controllers.Finally,the stability and convergence of the system are proved,and simulation results validate the effectiveness of proposed controllers.

关 键 词：非线性 多模型 自适应 零阶接近有界 全局收敛性 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]