基于正交小波网络的非线性系统辨识新方法

A NOVEL METHOD FOR IDENTIFYING NONLINEAR SYSTEMS USING ORTHOGONAL WAVELET NETWORKS

在非线性系统辨识中 ,系统输入数据往往是非均匀分布的。对于此类问题 ,正交小波网络处理起来较复杂。本文基于小波网格系 ,提出两种近似方法 ,可简单有效地解决这一问题。该法最大特点是计算简单 ,便于应用。最后用于辨识非线性系统 ,仿真结果验证了该方法的可行性和有效性。

In the identification of nonlinear system, the system inputs are usually non-uniformly distributed. The approach using orthogonal wavelet networks to resolve the problem is complex. This paper presents two approximative methods based on wavelet grid structure to handle this problem simply and effectively. The advantages of these methods are computationally efficient and easy to use. Finally, nonlinear systems are identified. The results of simulation show that the provided methods are effective and feasible.