Research On Trajectory Data Clustering Method Based On Spatio-temporal Correlation

In recent years,with the rapid development of urban economy and the continuous growth of traffic industry,traffic congestion has become a major problem affecting life of citizens and urban development.Through satellite,mobile phone and other positioning tools,the acquisition of spatio-temporal data is becoming stronger and stronger.How to mine the acquired trajectory data to provide the residents with fast and effective traffic information,so as to facilitate travel,slow down traffic congestion,and improve the ability of road passage,has become a subject of great interest to researchers.However,the trajectory data is far beyond the scope of human visual understanding,and has the characteristics of temporal attributes and spatial attributes,the traditional clustering method has some limitations in its analysis and processing.Aiming at the characteristics of position repetition and path repetition in trajectory data,this paper studies the trajectory clustering method based on association,and designs and completes the corresponding prototype system.The specific research work is as follows:Firstly,a trajectory data clustering algorithm based on spatio-temporal correlation between trajectories is proposed in this paper,which is based on the feature of a large number of geographical positions repetition of the trajectory clustering algorithm.This method mainly consists of two stages.In the first stage,the initial center representative point is first determined based on the minimum residence time limit and radius r,and then the maximum distance of the cluster is used as the radius R corresponding to the initial center representative point.Then,according to the minimum moving time constraint,the adjacent initial center representative points are merged and the radius R is adjusted to get the center representative point set.In the second stage,mainly deals with the new trajectory data.First of all,trajectory points is matched with the center representative point set,the successful points are deleted to generate a new trajectory,and then the first stage operation is performed on the new trajectory with clustering value.Finally,the center point set is updatedaccording to the clustering result,and the clustering is completed.Experimental results show that the algorithm can effectively reduce the time complexity of the algorithm.Second,a trajectory data clustering algorithm based on the spatio-temporal correlation and access factor between trajectories is proposed to solve the problem of a large number of iterations in a certain region.Firstly,the visit ratio VT and time t of each center point are calculated,and the center point with higher visit rate is selected to match first.Then the selected center point is matched by the introduced access factor AF and the matching sequence is put into the repeated path rout sequence.In the process of processing the new trajectory data,the NCR is matched according to the path in the rout sequence,and the remaining data points with clustering value are regenerated to the center point.Finally,if a new center point appears in the NCR,the rout sequence needs to be updated at the same time.The experimental results show that the introduction of access factor AF can alleviate the time waste caused by delayed matching repeat paths and improve the time efficiency of trajectory clustering.Thirdly,the prototype system of trajectory data clustering analysis based on spatio-temporal correlation is designed and implemented.The main functions of the prototype system include: central point set generation,repeat path generation,clustering result analysis,clustering results saving and other functional modules.The running results of the system show that the system can get the location of traffic congestion faster,so that the users can choose a fast and convenient traffic route as the basis.