Travel Pattern And Private Drivers’ Profile Based On License Plate Recognition Data

With the rapid development of China’s social economy and the steady improvement of people’s quality of life,the purchase of private cars has seen a sustained growth,and the road traffic flow has also increased,accompanied by increasingly serious traffic congestion,traffic accidents and social security problems.At the same time,the contradiction between the backward infrastructure construction and the growing demand for traffic management has become more severe.The construction and improvement of the road traffic bayonet monitoring system provides a basis for people to obtain urban traffic flow information and study the urban mobility of traffic demand.Exploring the profile of private car drivers can study the mobility of the city through the portrayal of traffic demand and travel characteristics,in the meanwhile,help the traffic management department to effectively control and supervise the traffic,and assist in the formulation and implementation of various traffic management decisions.This paper uses the data mining method to label the travel behavior of private car drivers based on the license plate recognition data in the road traffic bayonet monitoring system,drawing the driver’s travel portrait and depicts the driver’s travel mode.The main research contents of this paper are as follows:(1)A model of a driver’s travel profile is proposed.The concept of "user profile" in the field of e-commerce is extended to the field of transportation.Combining the characteristics of traffic travel,the basic attribute label and the space-time attribute label are defined,so as to portray the profile of the driver.(2)An extraction method of travel profile label is proposed.Travel profile labels can be divided into two categories by nature: basic label and traffic label.Tag extraction methods are divided into rule-based and data mining.The specific method is to use the motor vehicle registration form and the driver registration form to extract the driver’s basic attribute labels,such as gender and age.The AVI data is preprocessed to extract the driver’s travel chain,and the driver’s traffic attribute labels,such as travel distance,weekly travel frequency and travel range,are extracted in a rule-based way.By mining the AVI data of long time,the residence and work place are identified,and then the commuting distance and commuting time are calculated.(3)Studying the driver’s travel time and space mode.The driver’s travel demand data is extracted by using the license plate recognition data.The temporal and spatial patterns of the travel demand are mined by utilizing the LDA model.The K-Means algorithm is adopted to mine the patterns of bayonet flow variation during the day.At the same time,the traffic density map is applied to explore the time and space of interest in urban travel.This paper uses the AVI data spanning one week in the downtown area of Wuhan to conduct experiments.The results show that the AVI data is a good data source for mining urban travel patterns and establishing driver profile models.At the same time,it shows that the driver profile has a good application prospect in the study of Geocomputation for Social Sciences...