Fast Map Update Based On Trajectory Data

With the wide application of electronic maps in location services such as route plan-ning,navigation,transportation task dispatching,and vehicle scheduling,people have put forward higher requirements for the accuracy and real-time of map update.The phe-nomenon of missing roads,closure road,and deviation road in the map not only cause users to get lost and detour,but also cause traffic accidents and traffic congestion.There-fore,for the purpose of save travel costs and maintain normal traffic order,it is urgent to design an appropriate map update method.The popularization of intelligent travel soft-ware and the rapid development of Beidou navigation and positioning technology have derived large-scale trajectory data,which has the characteristics of large data volume,wide coverage,and rich temporal and spatial information,providing new ideas for rapid map update.The existing track-based map updating methods focus on the discovery of missing roads,pay little attention to the change of road capacity caused by road closure,and lack the timely calibration of offset roads in the road network,while closed roads and offset roads often bring great interference to map users during navigation.To solve the above problems,this paper proposes a fast map update framework based on vehicle trajectory.In order to improve the efficiency and effectiveness of map update,in the data preprocessing section,trajectory segmentation,noise filtering and trajectory reduction are performed on the trajectory data.At the same time,the relationship be-tween the context of the road relationship and correlation of the road in the road network data are obtained through map matching.In the road closure/opening detection section,the road network is divided into grids,the candidate enclosed areas are quickly located by observing the changes in the trajectory characteristics in each grid area,and the road closure/opening is identified according to the changes in road flow and the number of as-sociated road turns.On this basis,the closing/opening type is further determined by the duration and periodicity of the closing/opening behavior.The wrong road calibration sec-tion is obtained through trajectory clustering,centerline extraction and map matching,and then the map is quickly updated.Finally,three real trajectory data sets are used to verify the effectiveness of the proposed method.The main work of this paper includes the following aspects:· Data Preprocessing Aiming at the problem that trajectory data is easily affected by signal fluctuations and special driving behaviors,this paper designs a method to improve trajectory quality.First,the trajectory with abnormal temporal and spatial distribution in the trajectory data is segmented into multiple new sub-trajectories,and then the fluctuation points and self-intersections in the trajectory are detected by a sliding window model and removed as noise.Finally,the Douglas-Puckeral- gorithm is used to improve algorithm efficiency and realization of trajectory reduc-tion.In addition,this paper also uses map matching method to extract the context and correlation of the road.· Road Closure/Opening Detection This paper designs and implements a road clo-sure/opening detection method based on the exponential weighted moving average method.It first divides the detection area into grids of smaller granularity.Secondly,it uses the traffic flow data and the U-turn frequency data at the previous times of each grid unit combined with the exponential weighted moving average method to predict the current time value,and quickly locates the candidate closure area based on the obvious difference between predict value and the real value.Thirdly,it uses the map matching method to obtain the characteristics of significant changes in road flow and associated road turning frequency in the candidate closure area to screen out the closure/opening road.Finally,the closed/open roads are continuously moni-tored the closed behavior mode is further determined based on the duration and peri-odicity.By continuously monitoring the duration and periodicity of the closed/open channel,the closed behavior mode is determined.· Wrong Roads Calibration Wrong roads include missing roads and road deviation.In order to calibrate the wrong roads,this paper uses the trajectory clustering algo-rithm based on Hausdorff distance to obtain the trajectory clusters of different roads,and uses the force-attraction algorithm to adjust the trajectory clusters.Then,uses minimum sampling method to extract the road centerline.Finally,the map match-ing method is combined with the original road network to realize the calibration of the deviation roads.In summary,this paper designs a framework for fast map update based on trajec-tory data and designs corresponding algorithms.Experiments on synthetic and real-world network datasets show the effectiveness and efficiency of prorposed methods.