基于模糊模式识别的高速公路交通事件的自动检测  

A Fuzzy Pattern Recognition-based Approach to Automatic Freeway Incident Detection

作　　者：焦军彩[1] 唐健[2] 

机构地区：[1]东南大学数学系,江苏南京210018 [2]南京工业大学数学系,江苏南京210009

出　　处：《常州工学院学报》2006年第2期8-11,共4页Journal of Changzhou Institute of Technology

摘　　要：通过模糊模式识别的方法,对高速公路交通事件进行自动检测。根据交通流量、速度和占有率,构造不同交通流状态的隶属函数,根据最大隶属度原则进行交通流状态的识别;构造3个变量的增量关于事件和非事件的隶属函数,根据最大隶属度原则进行事件的模糊识别。对算法进行离线性能评价,结果表明,模糊模式识别方法优于一般的事件自动检测(AID)方法,从而为AID提供了一种更科学有效的方法。Incidents occurred in the freeways can be detected automatically by way of fuzzy pattern recognition. According to traffic flux, rate, occupancy, various subjection function of state of traffic flow was constituted and so was the subjection function related to incidents and non - incidents in light of the increment of three variables. According to the principles of maximum degree of subjection, the traffic state was recognized and incident was detected. Based on the above -mentioned principle of maximum degree of subjection together with off - line comments on calculation, it shows that fuzzy pattern recognition is superior to the regular way of traffic accidents automatic detection, therefore providing a more scientific and effective way.

关 键 词：模糊模式识别 事件自动检测 隶属函数 最大隶属度 

分 类 号：O159[理学—数学] TP213[理学—基础数学]