Study On The Method Of Analyzing The Driving Characteristics Of Private Cars Which Have Regularity In Urban Environment

With the continuous acceleration of industrialization and urbanization in China,the ownership of automotive vehicle in the urban area has increased rapidly,and the vehicles have become a key factor that affects and restricts the healthy development of urban traffic system.On the one hand,it brings great pressure to transportation,safe driving and environmental protection.On the other hand,with the development of information processing,data mining and other technologies,it also offers opportunities for the acquisition,analysis and processing of dynamic traffic trajectory data,as well as the development of related industries.The way to accurately collect and obtain the trajectory information of the private car,including the position and driving state of the vehicle,and analyze and excavate the characteristics of the trajectory and the moving pattern,can provide a new way to explore the dynamic changing trend and the evolution rule of the traffic flow.It is an important study of the application of the urban intelligent transportation system and the networks of vehicles.Therefore,it is very important to obtain the trajectory data and analyze the behavior characteristics of private cars.This paper mainly focuses on analyzing the driving characteristics based on the clustering method to help people understand the travel regularity of private cars and the influence on urban traffic flow of their travel behavior.The main work of this article is as follows:In this paper,the historical trajectory information of some private cars in Shenzhen is analyzed,and valuable data attributes are selected.In order to reduce storage space and improve the accuracy of the algorithm analysis and experiment,the original data obtained from vehicle's trajectories is preprocessed.Firstly,removing the invalid or redundant which is detected from trajectory based on the time and distance threshold.Secondly,using the Kalman filter to calibrate the trajectory.Finally,data before calibration and data calibration after calibration operation are compared..A TAD(Trajectory Aggregation Detection,TAD)algorithm which based on clustering is proposed to detect the regularity of vehicles for analyzing the driving characteristics of vehicle trajectory.The algorithm includes three steps: dynamic neighborhood adaptation,trajectory clustering,and understanding travel regularity.The positioning error of the vehicle is influenced by many factors,such as collecting equipment,environment and so on,therefore,using the dynamic neighborhood radius at this time can greatly improve the test accuracy.Secondly,the algorithm takes into account the temporal and spatial clustering of trajectories,rather than just spatial analysis.Finally,the definition and type of regularity are given.The detection results of TAD and two classical clustering algorithm were compared in the experiment.Furthermore,This paper is based on the trajectory data of 1000 private cars for one month.The results show that the proposed method can well identify the regularity of vehicle travel behavior.The TAD algorithm is applied to the vehicle regular trip detection prototype system.The system is mainly based on the B/S architecture.The user only needs to interact with the server in the web browser.According to the design requirements,the server uses the tornado framework for fast access.The system function module mainly provides the user with the platform which can input and upload the trajectory data to be detected.The related tutorial module gives an example of the detection,how to evaluate the performance of the system,explain the detection principle of the system and the source of the data source of the statement evaluation detection method.Finally,the system will run as part of the intelligent vehicle networking system.