Research On The Method Of Vehicle Travel Path Recognition Based On Discontinuous Electric Police Data

With the popularization of traffic information and the continuous development of data mining technology,it is possible to mine the data collected by the electronic police system to extract the complete path information of vehicles.The abundant traffic information contained in the electric police data is the most direct reflection of the urban travel law,and the related problems based on this are paid more and more attention by traffic researchers and managers.But at the same time,on the one hand,the coverage rate of electronic police system is low due to financial factors in some cities,on the other hand,the complex environmental factors cause the instability of the working condition of the electronic police system,which leads to the lack of data collected by the system,resulting in the "blind spot" of path recognition and the inability to directly obtain the complete travel path of vehicles.In view of this problem,the establishment of a set of path recognition method under the condition of discontinuous electric police data has practical significance for reducing the equipment layout requirements to save costs and improve the robustness of the electronic police data under the condition of meeting the accuracy requirements.Based on the stored license plate number,time and location information,this paper focuses on solving the problem of incomplete travel path identification caused by the lack of data.The research work of this paper are summarized as follows:(1)By consulting a large number of literature at home and abroad,this paper summarizes the application status of domestic and foreign electric police data and the status quo of missing path identification and repair,and summarizes the research status of the index weight calculation based on the chromatography analysis method and entropy weight method.(2)This paper analyzes the principle of data acquisition and the causes of abnormal data,and puts forward the detection process and processing method for each kind of abnormal data.In order to ensure the maximum amount of data,according to the space-time similarity of adjacent travel records and the fuzzy matching of license plate,the abnormal data is restored,and the utilization rate of data is improved.(3)The calculation method of travel time threshold of adjacent intersection is proposed,and vehicle path recognition and sub path cutting are completed based on the time threshold.The missing path is identified according to the topological relationship of road network.In order to narrow the scope of path search algorithm,Dijkstra algorithm and deletion algorithm are applied,and screening rules are formulated to search the candidate path set on the missing path.(4)This paper establishes a path recognition model based on discontinuous electric police data,discusses the travel rules of private cars and taxis,and determines the different indexes of the model;in order to facilitate the calculation,puts forward the quantitative formula of the index;uses the combination of analytic hierarchy process and entropy weight method to calculate the index weight,and increases the number of indexes after standardization in order to solve the 0,1 problem in the process of index standardization of entropy weight method Finally,based on the path recognition model,the path recognition of private cars and taxis are studied respectively.(5)Combined with the electric police data of Yicheng County,the example is verified.This paper analyzes the processing results of the electric police data,and constructs the road network topology model;selects a pair of electric police intersections to verify the calculation process and effectiveness of the model,and gives the conclusion that the path identified based on the path recognition model is consistent with the actual path;designs three groups of experiments with different path lengths and different number of missing points to verify the accuracy of the model,and gives some suggestions With the increase of the number of missing points in the long-distance path,the uncertainty of the model results increases,and the accuracy of path recognition decreases,but compared with the shortest path algorithm,it still has good performance;finally,the sensitivity of the index weight is analyzed,and the conclusion that the model results can maintain good stability in the reasonable range of the weight value is given Under the condition of sufficient accuracy requirements,the proposed method can be applied to identify the vehicle travel path under discontinuous electric police data.