论文部分内容阅读
模糊基函数网络能很好地以“if…then…”形式利用专家的知识，且以这种模型进行辨识，其辨识误差方程与辨识参数成线性关系，因而非常适用于实时辨识系统中．本文研究了这种网络模型辨识的持续激励条件和基于梯度法的辨识系统稳定性，给出了它们存在的充要条件．
Fuzzy basis function network can use expert knowledge well in the form of “if ... then ...” and recognize it with this model. Its identification error equation is linear with identification parameters, so it is very suitable for real-time identification system. In this paper, we study the continuous excitation conditions identified by this network model and the stability of the identification system based on gradient method, and give their necessary and sufficient conditions.