强噪声下已知信号的模糊神经网络识别

RECOGNITION OF KNOWN SIGNAL IN STRONG NOISE BASED ON FUZZY NEURAL NETWORKS

以计算机模拟为基础研究了用模糊神经网络方法对被噪声严重污染的已知信号进行识别的问题，研究表明，将模糊隶属函数和BP神经网络相结合对信噪比极低的信号有较强的识别能力，本文还从实用性角度讨论了这一识别方法的可行性，这为强噪声下的磁共振信号识别问题提供了新途径．

In this paper, problems associated to recognition of known signal submerged in noise is studied with computer simulations by Fuzzy Neural Networks. The research results show that, under very low signal-to-noise ratio, a very high recognition rate was still kept by the networkss with combination of fuzzy membership function and BP algorithm. In addition, the practicability of this recognition method was investigated.The approach opens a new way to recognition of signal embeded in strong noise in NMR.