基于Soft Fuzzy ARTMAP算法的地铁隧洞光纤光栅传感火灾识别研究*

为解决传统火灾报警系统在应对地铁火灾报警时存在的即时性差、灵敏度及智能化水平低等问题，基于Fuzzy ARTMAP网络，结合Yu范数相似度准则与软竞争学习机制，提出1种软竞争学习Fuzzy ARTMAP算法，弥补传统Fuzzy ARTMAP网络胜者为王规则导致的区域重叠而产生误判的不足；结合地铁光纤光栅传感网络数据，将该算法应用于地铁火灾识别。结果表明:与传统的Fuzzy ARTMAP相比，该算法可快速有效地识别地铁火灾趋势，为地铁火灾识别系统研究提供理论支持。

Study on fire detection by FBG sensing in subway tunnel based on algorithm of Soft Fuzzy ARTMAP

In order to overcome the problems of poor instantaneity,low sensitivity and low level of intelligence when coping with the subway fire alarm by using the traditional fire alarm system,a soft competitive learning Fuzzy ARTMAP (Soft Fuzzy ARTMAP) algorithm was proposed based on the fuzzy ARTMAP network combining with the similarity criterion of Yu norm and the soft competitive learning mechanism,and it maked up the shortcoming of the traditional Fuzzy ARTMAP network that the rule of “winner take all” caused the area overlap,which led to the misjudge.Combined with the data of subway FBG sensing network,the algorithm was applied in the subway fire detection.The results showed that compared with the traditional Fuzzy ARTMAP,the algorithm could identify the subway fire trend quickly and effectively,and provide theoretical support for the research of subway fire detection system.