Research On Vehicle Trajectory Based Urban Road Impedance Model And Path Recommendation Method

The efficient urban transportation system not only ensures that people can enjoy high-quality transportation services,but also promotes the development of the entire society and economy.With the advancement of urbanization,the size of the urban population is getting larger.Thanks to that,the contradiction between the transportation and road facilities is increasingly apparent,and urban transportation is facing more and more difficult problems and challenges.Excessive traffic flow causes severe overloading of roads,making existing transportation in cities chaotic and inefficient,which has caused huge economic losses to individual users and even the entire society.Therefore,how to alleviate the urban congestion has become the focus of research at home and abroad.In fact,the traffic flow is a collection of path selection behaviors of all travel users,so it is particularly important to study the path recommendation strategy in the transportation system.According to the survey,in addition to personal preferences,the user's path selection depends mainly on the path impedance.Consider this,this paper focuses on the research of urban road impedance models and path recommendation algorithms.The current research on the impedance model of urban roads is mainly based on the BPR road impedance model proposed by the U.S.Highway Administration.What's more,most impedance models are based on road flow data.In reality,the coverage of traffic road flow monitoring systems is very limited,which will make some uncovered roads difficult to estimate.Fortunately,the ubiquitous probe cars in the urban road network can provide massive GPS trajectory data,which includes a lot of road traffic information.In view of this,this paper proposes an urban road impedance model based on vehicle trajectory data.In addition,this article also analyzes the process of vehicles passing through road intersections,and considers the impact of vehicle queue delay time on the impedance model.By comparing the experimental results of multiple roads,we prove that the proposed urban road impedance model has good robustness while ensuring high prediction accuracy.With the absence of accurate path recommendations,users can only choose paths blindly based on their own experience and little road information.This can easily lead to the traffic congestion and reduce the overall efficiency of the system.Therefore,the traffic management department must coordinate the distribution of traffic flow,and continuously optimize the path recommendation strategy to maximize the utility of urban road facilities.The current dynamic optimal path recommendation methods are mostly based on the real-time road traffic information.However,on the one hand,the existing multi-source real-time road condition data does not have a unified fusion standard,resulting in a single and incomplete real-time data content.On the other hand,the coverage area of the real-time detection system of road traffic information in China is very limited,which makes the existing optimal path recommendation algorithms not applicable to the roads not covered by the real-time detection system.Based on this,this paper proposes a Q-learning based urban road path recommendation method.This method does not rely on real-time road traffic information,but uses a combination of historical vehicle data and real-time observation data of driving users to make up for the limitation of the real-time detection system,and the algorithm has fast convergence speed with ensuring the performance of the algorithm,i.e.,the time complexity is reduced.After comparing multiple path recommendation methods,it is proved that the effect of the path recommendation method based on Q-learning algorithm proposed in this paper is the closest one to the best path decision effect.