Research On Track Compression And Clustering Algorithm Based On Track Eigenvalue

With the popularity of mobile devices,location-based services are more and more popular,such as map navigation,Didi taxi and other positioning software,which are widely used in people's lives.A large number of trajectory data can help people to understand the real-time traffic situation,plan travel,and analyze the characteristics of crowd behavior.The rapid growth of track data brings huge pressure to the storage and network transmission of mobile devices.Track data compression technology is an effective way to solve the track data expansion.The existing methods extract feature points by threshold,which can not effectively reduce the compression error of trajectory.In this paper,a method based on maximum points is proposed to obtain the feature points in each track segment,so that each feature point is the track point with the largest amount of information in the track segment,which greatly reduces the error of the compressed track.The track data contains the spatial position,time,speed and other information of the track.In the segmented track clustering,the original track is divided into sub track segments by selecting the key points in the original track,and the track clustering is realized by clustering algorithm.The selection of key points in track segmentation clustering becomes the key of track clustering.In this paper,the key points in the original track are obtained by using the angle variation of track points,and the original track is divided into sub track segments.The clustering analysis of track set is realized by using clustering algorithm,which effectively improves the clustering effect of track.The main contributions of this paper are as follows:Research on track compression: traditional track compression methods obtain feature points by setting thresholds,but the feature points are not the track points with the largest amount of information in the track,so there is a large error between the compressed track data set and the original track.In this paper,an on-line trajectory compression(MP)method based on "maximum point" is proposed.In each track,there is a track point with the largest amount of information as the feature point to represent the track.By searching for the maximum value point,the information of thefeature point is maximized,so that the error between the compressed track and the original track is minimizedResearch on trajectory clustering: by obtaining the key points of the original trajectory,it can not only be applied to the trajectory compression,but also to the segmentation clustering of the trajectory.The track points in the track contain space information,time information,speed information and angle information.According to the change of the angle of the track points,this paper proposes a clustering algorithm of the track segments based on the change of the angle.By comparing the difference between the angle value of the current track point and the angle value of the previous track point and the threshold value,we can judge whether the current track point is a key point and realize the purpose of track segmentation.Then the DBSCAN clustering algorithm based on density is used to cluster the track segments,which improves the clustering effect of the track...