The Estimation Method Of Real-time Traffic Based On Connected Vehicle

A sharp increase in car ownership in China has led to urban traffic congestion,and the conventional way to upgrade roads and control the increasing of the number of vehicles cannot solve the problem from the perspective of long-term development.Using the existing advanced technology to analyze the traffic state of the roads,provides the best real-time traffic information,for achieving an easy-flowing vehicular traffic.This thesis studies a real-time traffic estimation architecture based on a connected vehicle and the realization of a real-time traffic sharing system.Taking vehicles in Wuhan city as the population for the study,this thesis studies the key technologies of real-time traffic estimation algorithm,vehicle data acquisition and applications to traffic estimation.The main methods of the study are outlined in the subsequent sections.Data acquisition system was divided into vehicle-mounted data acquisition platform and a mobile data acquisition platform.The vehicle-mounted data acquisition platform involved the communication between the onboard computer and the vehicle network to collect the real-time speed and other data via the OBD-II interfaces.The mobile data acquisition platform comprised a vehicular communication device realized the communication between the mobile terminals and the vehicle network,with the collected data being uploaded onto a server.Real-time traffic estimation algorithm efficiency is an important factor in with regard to the impact of the data application.Firstly,vehicle data collected could be preprocessed.Due to different conditions,the relationship between road travel speed and vehicle speeds is more complex.This thesis used the average speed of different types of vehicles and the proportion of these types as input variable,with the road travel speed as the output for constructing an RBF neutral network and a support vector machine.Performance indicators in the form MSE and MAPE were used for evaluating the performance of the RBF neutral network and the support vector machine.On the average,the MAPE of the support vector machine was 6.95% less than the RBF neutral network.Comparing the estimated results of the two algorithms indicated that the support vector machine was more reliable and accurate.Therefore,the result of the support vector machine was used as road travel speed of the target segment.The estimated results of support vector machine can serve the public by being able to predict the traffic situation.This thesis achieved a traffic display system based on the terminal platform and the Baidu map.The traffic state of urban road was divided into three levels,which were flow,amble and congestion,to provide simple and intuitive traffic conditions of hot spots around the city of Wuhan.Additionally,this thesis achieved real-time location and historical itinerary playback functions with GPS technology.The thesis developed the real-time traffic estimation platform,including vehicle data acquisition,data management and traffic display,and has a certain value for real-time traffic study.