Study On Monitoring Model Of Highway Congestion Based On Mobile User Information

With the rapid growth of car ownership in China, highway congestion has become increasingly serious. How to monitor the congestion effectively is the foundation of managing and controlling highway traffic. Currently, the main monitoring technologies include fixed monitoring like ring induction coil and GPS floating vehicles monitoring. However, there are some problems like narrow coverage, high cost and so on. Therefore, in order to monitor highway congestion with low-cost, all-weather, large-scale and real-time, mobile user information is used to collect highway traffic information and identify congestion.Firstly, the thesis studies and analyses the data characteristics of mobile communication network user information as well as the highway traffic characteristics. According to the classification of highway service level in China, the thesis mainly studies the sensitivity of highway vehicles' speed, traffic density and traffic volume to the change of traffic state. And then, average travel speed and traffic density are selected as the measure index of highway congestion.Secondly, the overall scheme of highway congestion model is designed based on the selection of congestion metrics and the demand analysis of congestion model. By analysing and studying the characteristics of wireless coverage of highway mobile network, the spatial matching of highway and mobile network is completed, and a way is proposed to divide highway into segments dynamically based on the active degree of matching projection points. Combining user's moving trajectory extracted from the mobile user information with traffic characteristics of highway, a method is put forward for recognising vehicle-mounted mobile phone users on the highway. By analysing and studying common calculating methods of speed and traffic density, the traffic density is calculated according to the average number of phones in-vehicle and the average speed of road section is calculated by using speed weighted fusion. In this way, the accuracy of calculation results is effectively improved.Thirdly, aiming at the randomness and fuzziness of traditional digital classification for traffic state, a cloud matching method is used to recognizing highway congestion. A similarity matching method based on overlap of cloud images is proposed. And the validity and reliability of the algorithm are verified by experiments.Finally, the model is verified by using the mobile user information provided by a Mobile Corporation on the data processing platform. Through comparing the results of the model with the data of fixed detector on the highway, it shows that the congestion model proposed in this thesis can meet the design requirements of the model in terms of speed, density and congestion identification, which has some practical value.