Research On Map Matching Method Based On Vehicle GPS Trajectory And Mobile Signaling

In recent years,while urban development has brought us convenient modern life,there are many problems and challenges,such as traffic congestion,exhaust pollution,backward urban planning and so on.On the other hand,urban computing has become a research hotspot in recent years with artificial intelligence,big data mining and other technologies playing an increasingly important role in urban planning,traffic supervision,energy allocation and other fields.Map matching,as an important part of urban computing,has a wide range of applications in path planning,navigation,road traffic regulation and so on.Map matching is a method of converting the original traffic trajectory points into sections in the traffic network.This point-to-line transformation can provide high quality data sources for subsequent scientific research related to road flow and network optimization.In view of the importance of map matching mentioned above,this paper focuses on optimizing the map matching method for urban traffic trajectory.In the problem of map matching,the GPS trajectory of coarse-grained sampling is more difficult to match than that of fine-grained sampling,and the accuracy of general incremental algorithm is worse than that of global algorithm,but the incremental algorithm is faster.Therefore,an incremental map matching algorithm based on weighted shortest path is proposed to solve the problem that incremental algorithm is ineffective in matching coarse-grained GPS trajectory.The traditional incremental algorithm only considers the spatial relationship between sampling points and target sections when matching.The proposed matching algorithm takes into account the spatial relationship,the connection between front and back sections,the properties of sections,people's driving habits,speed limitations and other factors.Firstly,it improves the matching accuracy of single GPS points between candidate sections,and secondly,it improves the matching accuracy of filling the gap between the two matched road sections.Experiments show that this algorithm not only guarantees the good time efficiency of incremental map matching algorithm,but also improves the accuracy of matching results.Vehicle GPS trajectory has some limitations,such as high cost of acquisition,high energy consumption,low coverage of road traffic,and mobile signaling data almost cover the behavior trajectory of all the population in the city.Therefore,this paper proposes a map matching algorithm based on historical experience and hidden Markov model to match mobile signaling trajectory to traffic network.Mobile signaling data has many noise points and poor quality.Trajectories of various travel modes are mixed together.First,data preprocessing is used to smooth positioning errors and filter noise.Then,trajectory pattern recognition is carried out to distinguish vehicle trajectories from other types of trajectories.In the matching stage,in order to solve the problem of sparse sampling of mobile signaling trajectory,this paper proposes a trajectory interpolation method based on historical trajectory path and time interval factors,and constructs a probability-based hidden Markov model according to the input trajectory and road network data,which ingeniously transforms the problem of map matching into the problem of Viterbi decoding.Experiments show that the matching algorithm based on historical experience and Hidden Markov Model for mobile phone signaling trajectory map is more accurate than the traditional algorithm using only hidden Markov model.