Traffic Tharacteristics Mining Based On RFID Data Of Vehicle

With the rapid development of information technology,urban traffic data acquisition technology and means gradually increased,more and more easy to obtain a large number of urban traffic information.Looking for the urban traffic characteristic from these massive data is one of the important fields of the intelligent traffic development,it provides an effective way to solve the traffic problem which is caused by the clustering effect and the scale economy in the urbanization process.Based on the traffic data collected by the RFID technology,this thesis explores the urban traffic characteristics and reveals the current rules of urban transport operation.Through these studies can provide the strong basis for the urban traffic fine management,simultaneously to provide the important reference value to the urban transportation policy formulation.Therefore,the traffic characteristics based on the vehicle RFID Data mining has a certain social and economic significance.In this dissertation,the author have in-depth analyzed traffic characteristics based on RFID technology,Hadoop platform and other technologies,By studying the data mining methods of DBSCAN density clustering and Apriori Association rules,this thesis puts forward four kinds of methods for determining and identifying traffic characteristics,and applies these methods to the mega-city road network.Firstly,in order to improve the precision of traffic characteristics of massive data mining,the dissertation divides the characteristics of the field attribute in the acquired RFID data into three kinds,and uses the corresponding method to deal with it.Secondly,in order to find the traffic characteristics,such as determining the number of daily trips,the identification mode of the commuter vehicle,the commuting track of the commuter vehicle and the key sections of the road network,and tracing the direction of the traffic flow sources in the key sections,This paper presents a method of vehicle trip number recognition based on DBSCAN density clustering,a commuter vehicle identification model based on DBSCAN density clustering,a commuter travel trajectory feature mining algorithm based on Apriori Association rules and a method for identifying traffic flow source direction based on Apriori Association rules for road network key sections;Then,in order to improve the efficiency of data mining,this paper focuses on the large data mining platform Hadoop,and compares the efficiency of the Hadoop platform and Oracle database in data storage and processing,and finally,using the historical RFID data quantity(about 80GB)of the Chongqing in the main urban area,The DBSCAN density clustering and Apriori Association rules algorithm are used to excavate the traffic characteristics of Chongqing main urban area in the Hadoop.Through a month’s traffic in the main city of Chongqing.The experiment of RFID data mining shows that there are some regularity characteristics of vehicle travel in main urban area,such as daily travel times,commuter number,commuting track of commuter vehicle and key sections of urban road network.By analyzing the characteristics of these laws,we can provide suggestions for the fine management of traffic in the main urban area of Chongqing,and provide reference for the development of traffic policy in Chongqing.