Design And Implementation Of Expressway Traffic Flow Analysis System

With the rapid development of expressways,more and more people choose the method of the expressway,and thus there is a phenomenon that the toll stations are always not accessible.In order to ensure that people's travel safety,traffic management department proposes transform from highway information system to intelligent transportation,so that they could control and manage vehicle traffic through technical means better.Therefore,the project team decides to develop this system,which could reveal the interaction between the regions by analyzing the traffic flow,which could provide a reliable data basis for traffic guidance and diversion to reduce traffic congestion,and which could predict the traffic conditions of the expressway more accurately through providing traffic data for travelers.Thereby,people will determine the best travelling time according to these data.This article first introduces the background of the project and the development both at home and abroad and analyzes the functional and non-functional requirements of the entire system.According to the needs analysis,the author designs the system architecture,divides the functional modules and designs the system database.In addition,the author designs and implements the system in detail.Above all,the author independently completes the traffic flow prediction model based on Long Short-Term Memory(LSTM)and the vehicle pairing algorithm based on Naive Bayes algorithm,as well as the development of the system's individual vehicle flow distribution,overall vehicle flow distribution,list ranking and traffic flow prediction module.In the process of system design and implementation,at first,the author introduces the design and training of two algorithms.For one thing,the traffic flow prediction algorithm based on LSTM,which fully considers three time-dependent factors and geographical factor of traffic flow,to make the accuracy of prediction error rate is 10.05%,the result is that the model performs better than the Support Vector Regression prediction model.For another,the function of vehicle pairing algorithm based on the Naive Bayes algorithm,is to match with the information except the license plate number and return to the original travel traces,when the stored data has the wrong and the vehicle will cross the province.The matching accuracy rate is 84.6%.Secondly,the author introduces the design and implementation of each module of the system,including individual vehicle flow distribution module,overall vehicle flow distribution module,list ranking module,traffic flow prediction module.The module of individual vehicle flow distribution shows the details of the traffic of individual vehicles and analyzes the individual traffic flow from the individual perspective,then counts comprehensively the traffic information of vehicles on the expressway,and thus providing basic support for the overall vehicle flow distribution module.The module of overall vehicle flow distribution acquaints the overall flow distribution of passenger flow and cargo flow,reflecting macroscopically the interaction between regions.The module of list ranking collects the vehicle traffic characteristics toll stations and sections which reflect the overall operational characteristics of the expressway.The module of traffic flow forecasting analyzes quantitatively the toll stations of top ranking soon after,and then system could predict the short-term traffic operation in the region soon after.After the latest testing,the system can meet the demands.At present,the system has been put into use,and the system is running normally.Furthermore,the function is stable,and the expected goal is achieved.