非线性系统的神经网络自学习控制

提出了一种对非线性系统的神经网络自学习控制方法，基于逆动力学控制的思想，构造了神经网络结构一致的控制器和辩识器。辨识器采用多层前向网络结构和广义Ｄｅｌｔａ学习规则算法实现了对系统逆动力学模型的动态辨识，并通过在线动态传递权值给神经网络控制器的方法实现了神经网络辨识器的神经网络控制器的有机结合，从而使整个控制系统具有很强的自适应和自学习能力，所提出的控制方案可适用于不含滞后环节和包含滞后环节的非线性

Self learning Control of Nonlinear System Based on Neural Networks

A method of self learning control based on neural networks for a nonlinear system is proposed.Based on the thought of inverse system control,the neural networks structures of the controller and the estimator are the same.The system estimator based on multi layer feedforward networks which applies Delta rule algorithm realizes the identification of the inverse dynamic system model.And by on line delivering weights matrix to network controller, the estimator realizes the connection with the system controller .The whole control system has the adaptive and the self learning abilities.The method in this paper can be applied to time delay or non time delay nonlinear systems. Its effectiveness is shown by two simulation examples.