具有不确定性的非线性系统自适应神经网络L_2增益控制  

An adaptive neural network L_2-gain controller for nonlinear systems with uncertainty

作　　者：李桂英[1] 魏莹[1] 张扬[1] 孙来军[1] 

机构地区：[1]黑龙江大学电气工程及其自动化实验室,黑龙江哈尔滨150080

出　　处：《智能系统学报》2009年第4期357-362,共6页CAAI Transactions on Intelligent Systems

基　　金：黑龙江省自然科学基金资助项目(F2007-07);黑龙江大学青年科学基金资助项目(QL200736)

摘　　要：针对存在不确定性的非线性系统,提出了自适应神经网络L2增益控制器设计方法,将基于Hamilton-Jacobi-Issacs(HJI)不等式和自适应神经网络策略相结合,有效地克服了需要被控对象精确建模的局限性.神经网络对系统模型的偏差进行拟合;为了补偿拟合误差,引入补偿控制器和神经网络权值自适应调节律,通过在线自适应修正神经网络权值,来保证闭环系统满足相应的L2性能准则.仿真结果表明提出的控制器设计方法是有效的,克服了一般方法需要被控对象精确建模的局限性.A scheme for an adaptive neural network L2- gain controller was proposed for nonlinear systems with uncertainty. By combining the Hamilton-Jacobi-Issacs （HJI） inequality with an adaptive neural network, limitations on the precision of previous models can be effectively overcome. With this controller, errors from the model were fitted by the neural network. In order to compensate for the fitting errors, a compensation controller and an adaptive law for the weights of the neural network were introduced. By on-line adaptive adjustment of these weights, L2- gain perfomlance of the closed-loop system could be guaranteed. Simulation results are shown to demonstrate the effectiveness and the advantages of the proposed approach. To avoid the limitation of the precision model of the plant in the common approach.

关 键 词：L2增益 神经网络控制 HJI不等式 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]