论文部分内容阅读
为了实现模糊神经网络结构和参数的同时调整,提出一种基于无迹卡尔曼滤波(UKF)的增长型模糊神经网络(UKF-GFNN).首先,利用UKF对模糊神经网络的参数进行调整;然后,设计一种基于隐含层神经元输出强度的模糊规则增长机制,实现模糊神经网络的结构增长;最后,将所提出的增长型模糊神经网络应用于非线性系统建模.实验结果显示,基于UKF的增长型模糊神经网络能够实现结构和参数的自校正,并且具有较高的建模精度.
In order to realize the simultaneous adjustment of the structure and parameters of the fuzzy neural network, an improved UKF-GFNN (UKF-GFNN) is proposed.First, the parameters of the fuzzy neural network are adjusted by UKF; then , A fuzzy rule growth mechanism based on the output intensity of hidden layer neurons is designed to realize the structural growth of the fuzzy neural network. Finally, the proposed growth fuzzy neural network is applied to the nonlinear system modeling. The experimental results show that, UKF’s growth-type fuzzy neural network enables self-tuning of structures and parameters with high modeling accuracy.