摘要
该文从基本的智能技术——神经网络(NN)和模糊系统(FS)技术出发,探讨了神经网络与模糊系统相结合的基本理论,提出了一种基于模糊RBF神经网络的非线性滤波的方法。该方法将模糊逻辑的知识表达以及推理能力和RBF网络的快速学习和泛化能力结合起来,网络结构参数可按实际问题调整,对信号中有色噪声进行较高精度的逼近,来达到非线性滤波的目的。该滤波方法显示出很强的处理问题的能力,学习速度快,仿真结果表明了这种方法的有效性和可性行。
n this paper, the basic intelligent control technology——neural network （NN）and fuzzy system（FS） is introduced. Theories of neural network combined with fuzzy system are discussed. A method of nonlinear filtering based fuzzy radial basis function （RBF） neural network is proposed and the knowledge expression of fuzzy logic is combined with fast learning speed and generalization of RBF neural network. Its structure and parameters can be adjusted according to the real problem and filter nonlinear noise, for the better performance of the fuzzy RBF neural network in approximating. The way displays better ability, because of its advantages, its fast learning speed and simulation results testified the feasibility and validity of the proposed method.
出处
《计算机仿真》
CSCD
2006年第3期133-136,共4页
Computer Simulation
关键词
模糊
径向基函数
神经网络
非线性滤波
Fuzzy
Radial basis function
Neural network
Nonlinear filtering
