Research And Application Of T City Traffic Information Sharing Platform

With the rapid development of China's economy,traffic problems have gradually become the focus of attention.How to use modern technology to strengthen traffic management has become an important issue in urban development.Intelligent transportation system is an important solution to improve the management and control capabilities of traffic.Traffic information sharing platform is an indispensable platform for intelligent transportation system and plays an important role as a transportation hub in intelligent transportation system.Based on the current traffic situation in T City,this thesis aims at the main problem of urban transportation,based on the architecture model of SOA,and realizes the traffic information sharing platform through the Dubbo+SSM framework.At the same time,the traffic flow prediction model is studied by constructing the gradient lifting tree model,and the research results are applied to the construction process of the platform.The main research work of the thesis is as follows:(1)Research on platform construction technology and traffic wander prediction modelAfter researching the information sharing platform,the architecture and implementation technology suitable for the platform were found.And the SOA framework,Dubbo framework and SSM framework were introduced.At the same time,two prediction algorithms,ARIMA time series algorithm and XGBOOST gradient lifting tree algorithm are summarized.(2)T-city traffic status survey and platform demand analysisBased on the specific investigation of T City and the reference to a large number of documents,the traffic problems in T City were analyzed in detail compared with other successful cases.Based on the analysis of traffic problems,the users of the platform,the different needs of each user,the functions to be completed by the platform as a whole,and the non-functional requirements of the platform are analyzed.(3)Design and implementation of traffic information sharing platformAccording to the demand analysis of the traffic information sharing platform,the architecture of the platform and the network structure are designed.Based on the top-level design,the detailed design of the platform is divided into logical structure design and database design.In the logical structure design,the logical structure is divided into five modules,an information interaction module,a database management module,an information release module,an information processing module,and a platform management module.In the database design,the main discussion is that the database is divided into two types:distributed database and centrally managed data warehouse.We connect the database of the original subsystem through the network to form a distributed database,and then build the data warehouse by storing the historical data of the subsystem according to the business category.(4)Construct a traffic flow prediction model and optimize itThe thesis explains the ways to obtain data,introduces the specific structure of the data,and preprocesses the data.The ARIMA time series algorithm model established first predicts the data,and the ARIMA algorithm is found to be insufficient.The ARIMA time series algorithm is used as a feature of the XGBOOST gradient lifting tree algorithm,and the XGBOOST algorithm is optimized using three major types of features.Finally,the model prediction results and the time used in the prediction are compared and evaluated.The improved XGBOOST algorithm is more accurate and the running time is shorter.