Research On Reconstruction Of Individual Travel Path Chain Based On Repair Of License Plate Recognition Data

In recent years,with the continuous increase of urban vehicle ownership,the problem of traffic congestion has become increasingly severe,which has greatly affected the daily travel of urban vehicle-individuals.The individual travel chain contains rich traffic flow microparameters.Using travel path chains for data clustering,mining,and collision analysis to systematically and comprehensively evaluate the travel pattern and path selection behavior of vehicles in the urban transportation network,providing a strong support for coordinating the distribution of road network congestion and improving road vehicle traffic efficiency.Therefore,this paper focuses on the research of individual travel chain reconstruction,and explores the field of traffic flow restoration and route travel time estimation to promote the analysis of travel feature structure.Research on data preprocessing,data quality evaluation and data spatio-temporal correlation analysis based on license plate recognition data.The basic working principles and common layout strategies of smart bayonet equipment are introduced.The characteristics of license plate recognition data and perform preprocessing and quality analysis based on such data are explained.In addition,the method of screening and extraction of travel time samples at the road segment level and combined extraction of individual travel path chains are also introduced.At last,it describes the hidden spatio-temporal correlations to provide reliable support for subsequent research.Research on the traffic flow restoration model based on graph recurrence convolution generate adversarial networks.Modeling based on the lack of intersection traffic data monitored by high-definition smart bayonet of urban road network as the research background.First,construct three types of traffic flow datasets: random missing,non-random missing,and combined missing.Then,use the Tensorflow deep learning framework to build a graph recurrence convolution generation adversarial network(GRCGAN).Unit to effectively extract the spatiotemporal correlation of road network intersections,and finally compare the tensor decomposition algorithm of GRCGAN with BGCP?Ha LRTC and the GAIN algorithm model.The results show that the traffic flow restoration model proposed in this paper has better performance,which provides a basis for the calculation of road traffic density on travel chain reconstruction.Research on the estimation of route travel time distribution under the urban road network.Firstly,the Bayesian Information Criteria(BIC)is used to determine the numbers of traffic state clusters.In order to mine the spatial correlation of the travel time of adjacent road segments,GMM clustering is used to identify the traffic state;then,based on the Markov chain theory,the initial calculation The probability and state transition matrix represents the dynamic trend of the travel time of the link,and calculates the link connection probability to connect the information of the adjacent links;finally,it calculates the conditional path probability of each path in the road network,calculates the travel time distribution of each road segment,and calculates the final path travel time distribution by weighting the convolution calculation results of each distribution with the conditional path probability.The experimental results show that the improved conditional path probability weighted estimation method the paper proposes is more accurate and robust..Research on the urban individual travel chain reconstruction algorithm based on Gradient Boosting Decision Tree(GBDT)algorithm.Through the results of traffic flow restoration and route travel time distribution estimation research.Vehicles are matched by license plate number,and the corresponding travel chains sorted by timestamp are initially extracted and split according to the intersection adjacency matrix,estimated link travel time and estimated route travel time.Subsequently,the key impacted variables of vehicle route choice are identified based on travel behavior analysis and traffic conditions,and then a reconstruction method for missing path is developed based on GBDT.Taking the field LPN data from Nanming District of a Chinese city as an example,the proposed method in this study,existing traditional algorithms and machine learning algorithms are verified based on the accuracy and calculation efficiency.The result showed that the proposed method can achieve a high reconstruction accuracy of 91%,which has greater advantages in urban vehicle travel chain reconstruction than the other methods.