Type-2 Fuzzy Analysis Of Traffic Flow For City Main Road

With the continuous development of China's economy,the number of motor vehicles has increased dramatically,so the modern intelligent transportation system is also developing rapidly.How to effectively deal with the non-linear,time-varying and sudden traffic flow data and analyze the traffic flow is one of the important technologies to achieve traffic control and guidance.Most of the methods are through the shallow model in learning,it is very easy to fall into the local extreme value which can not be more complex mathematical operationsAccording to the variation of traffic flow for main roads,in the paper interval type-2 fuzzy sets are used to analyze traffic flow.The main contents of the paper are as follows(1)In order to improve the quality and representative significance of the data,the methods of stratified sampling and k-means clustering are combined to deal with the data for every direction of the main road.Because the traffic flow of weekdays and weekends changes greatly,and even in the same day,traffic flow also change greatly in different time period.Thus,in the paper,one day is divided into six periods,i.e.morning,forenoon,noon,afternoon,night,and for every period,the data of traffic flow is collected every five minutes.K-means clustering algorithm is used to cluster data every hour,and the sampling proportion of each time period is obtained,so as to obtain the sampling data.(2)Besides introducing the new definition type-2 fuzzy sets,the fundamental concept?representation and the operations of union,intersection and complementary are given.At the same time,the corresponding examples of operations are provided.(3)Central limit theorem is used to convert the dispersed point data into interval data,then interval type-2 fuzzy sets and confidence interval are composed to deal with the data of traffic flow.At the same time,type-2 fuzzy controller is adopted to analyze the compute the data,and the results are defuzzified,and then the conclusion of the paper is obtainedThe simulation results and error analysis show the feasibility and practicability of the method.The error comparison among the analysis methods of regression tree,BP neural network which has lower error result and higher fault tolerance rate.Therefore,the type-2 fuzzy analysis method adopted which has better effects.