基于神经网络的一类非线性系统参数估计

PARAMETER ESTIMATION METHOD FOR A CLASS OF NONLINEAR SYSTEMS USING NEURAL NETWORKS

基于Ｈｏｐｆｉｅｌｄ神经网络的神经计算原理，提出了一类非线性系统的参数估计方法：首先将系统参数估计问题转化为以系统模型残差平方和为目标函数的优化计算问题，然后利用连续Ｈｏｐｆｉｅｌｄ神经网络的神经计算估计待辨识参数．数字仿真结果表明，该方法不仅算法简单，而且估计精度高．

Based on the neural-net computing theory of Hopfield neural networks,theparameter estimation method is developed for a class of nonlinear systems in this paper. By usingthis method, the parameter eshmahng problem is transformed into the optimizing problem whoseobjective function is the mean square errors of the system outputs. Then the unknown parameteris estimated through the neural-net computation. The digital simulation has shown that the present approach is simpler and the estimating accuracy is better than other methods.