神经网络在非线性系统参数辨识中应用

要对非线性系统进行控制，必须掌握其模型．介绍一种辨识非线性系统模型的方法，该方法利用多层神经网络可以逼近任何非线性函数这一理论，提出针对非线性系统构造神经网络模型，并给出一种实用的BP算法．最后通过仿真实验，证明该方法切实可行．

The application of nervous network in identifying the non-linear system parameters

To control non-linear system, we must master its model first. This essay introduces a method of discriminating the model of non-linear system. According to the theory that the multilayer nervous network can approach each non-linear function. Two steps about how to master its model are suggested. The first step is to construct a nervous network in accordance with non-linear system. The second step is to give a practical BP algorithm. In the end of the essay, it is proved that this method is practical and effective by simulational experiments.