Research On Identification And Prediction Of Road Traffic Status Based On Mobilephone Signaling

The rapid increase in the number of cars has brought a series of traffic problems,such as traffic accidents,traffic congestion,traffic pollution and so on.Intelligent Transportation Systems(ITS)is an effective method to solve traffic congestion and reduce the occurrence of traffic accidents.Road traffic state recognition and prediction is an important part of intelligent transportation.How to identify and predict traffic status quickly,accurately and cost is an urgent problem to be solved in the implementation of intelligent transportation system.Based on the national development and Reform Commission funded by the mobile Internet and the 4G mobile communication(TD-LTE)industrialization special fund([2014]2328)and Key breakthrough fund funded project(2011A011305001)in key areas of Guangdong and Hong Kong,this paper has carried out the research on the road traffic status identification and prediction evaluation technology.The road traffic status recognition and prediction optimization scheme based on mobile phone signaling is put forward.It provides a scientific and reliable way to identify and predict the road traffic status under the influence of many environmental factors,and realizes the wide area traffic state recognition and prediction.In this paper,the key technical problems of current traffic state recognition and prediction are studied in depth and this paper tries to find a way to solve practical problems from the perspective of engineering application,and obtain the following innovative results:1.A mobile user location method based on the combination of 3D ray propagation and depth learning is proposed.This method solves the problem of poor simulation effect and large workload of manual collection,and improves the accuracy,location accuracy and efficiency of location fingerprint library.It breaks through the key technologies of radio propagation model correction,neural network kernel function improvement,communication fingerprint library sparse index,communication parameter threshold setting and fingerprint matching algorithm selection.It realizes the precise location of mobile users in the wide area,forms the location fingerprint Library Based on the corrected mobile communication parameters,and achieves the mobile user localization technology solution by matching fingerprint information.2.the algorithm of mobile user's trip pattern recognition is put forward,which solves the problem of instability and dynamism of mobile user's travel mode and the problem of travel mode recognition easily affected by environmental factors.It breaks through the key technologies such as dynamic threshold setting of mobile user travel pattern recognition,fast setting of DBSCAN clustering parameters and rapid clustering of mass data.By absorbing the advantages of BIRCH and DBSCAN,the fast clustering is realized by combining the two methods,and the fast recognition of the travel mode of mass mobile users is realized.A technical solution for adaptive recognition of mobile user travel mode based on density and dynamic threshold clustering is formed.3.A method of road traffic status recognition based on fuzzy comprehensive evaluation is proposed.It is difficult to solve the uncertainty of road traffic state discrimination and difficult to quantify.Based on the reasonable selection of traffic flow characteristic parameters,the membership function of road traffic state is applied to calculate the occurrence probability of various traffic states.Then,combined with the weight vector of traffic flow parameters,the recognition of road traffic status is realized.4.A road traffic state prediction method based on random forest is proposed to solve the problem of "dimension disaster" of traffic state prediction in complex environment.Three improvements are made to the random forest algorithm,which improves the accuracy and speed of the prediction.Using Hadoop's own model to quickly model,and then use spark tools to process distributed data of real-time flow data,and achieve real-time prediction of traffic state under multiple factors.These achievements have been applied in the 2015 iteration research and development project of China Unicom big data incubator(based on the flow path service and technology development of the industry),and achieved good results.This shows that the results of this paper have made a breakthrough in mobile user positioning,mobile user travel mode recognition,road traffic status recognition and prediction.The road traffic state recognition and prediction method based on mobile phone signaling extends the application scope of mobile user communication data to a certain extent,and provides effective means for the development of traffic information,and has broad application prospects.