Research On Anomaly Detection Of User Trajectory For Mobile Terminal Equipment

With the rapid development of wireless sensor networks,digital communication technologies,positioning technologies,and large-scale applications of mobile terminal devices,geographic location information generated by people's daily activities can be easily obtained,thereby providing powerful support conditions for trajectory data mining.Trajectory anomaly detection is an important branch of the research field of trajectory data mining.It can determine whether the moving object deviates from the normal trajectory route or whether the moving object has sudden behavior.If the abnormal trajectory of the moving object can be detected and the cause of the abnormal occurrence can be analyzed,then in the future,reasonable decisions can be made to avoid similar problems,and it can also provide clues to the public security department in detecting cases.This thesis focuses on a large amount of user trajectory information in mobile terminal devices,and mainly studies how to detect the abnormal trajectories of the user.In the abnormal detection of stay trajectories,in order to reduce interference from noise points to detect more stay points,a two-stage stay point extraction algorithm TSPEA is proposed;an abnormal stay area detection algorithm DAASRIAN based on improved agglomeration hierarchical clustering is proposed,which is used to detect the abnormal stay trajectories of users.In the abnormal detection of move trajectories,the entire trajectory is segmented by the extracted stay points,and the distances between the move trajectories are calculated from multiple dimensions.In order to reduce the influence of predetermined parameters on the detection of abnormal move trajectories,an algorithm for detecting abnormal movetrajectories based on OPTICS is proposed.Experiments using the Geo Life dataset show that the algorithm proposed in this thesis can extract more stay points and effectively detect the user's abnormal trajectories.In addition,this thesis designs and implements a user trajectory abnormality detection system,which can collect the user's trajectory from the mobile terminal device,find the user's stay points,and detect the user's abnormal stay trajectories and move trajectories.