摘要： 提出了一种分布式神经模糊网络和自学习模糊控制器的构成方法。它是模型的一种扩展,使其能进行模糊推理和构成自学习CMAC的模糊控制器。该方法除具有优点外,还具有以下特点:输入数据通过模糊划分和隶属函数后自动编码,对精度没有限制;从现场数CMAC据直接获取控制规则,即使对未训练的数据,也能结合插值和泛化两种能力,推理给出合适的输出。学习实例证明了方法的有效性。

关键词: 分布式神经网络, 自学习模糊控制器, CMAC

Abstract: This paper presents a distributed neural fuzzy network(DNFN), which is capable of extracting expert knowledge automatically, the network is an extended CMAC model in the fuzzy research areas. The input data is encoded automatically by means of the fuzzy sets, and the fuzzy interpolation mechanism is joined up with the generalization ability of CMAC in a DNFN. The effectiveness of DNFN is shown by numerical simulation results.

中图分类号: 

• TP183