Abnormal trajectory detection method based on fuzzy theory

Abstract

The invention is applicable to the technical field of trajectory detection, and provides an abnormal trajectory detection method based on a fuzzy theory, which comprises the following steps: S1, putting a trajectory with a fixed starting point and a fixed ending point into a trajectory set, and dividing the trajectory STi in the trajectory set into a plurality of trajectory segments STin based on a sliding window; S2, detecting the types of all track segments STin in the current sliding window, wherein the track segments include an absolute normal track segment, a local abnormal track segment and a normal track segment; and S3, carrying out statistics on the number of local abnormal trajectory segments and the sum of the number of normal trajectory segments and the number of absolute normal trajectory segments on the trajectory STi, further calculating an abnormal score of the trajectory STi, and if the abnormal score is greater than a set threshold value, determining that the corresponding trajectory STi is a global abnormal trajectory, otherwise, determining that the corresponding trajectory STi is a normal trajectory. According to the method, the anomalies are classified into local anomalies and global anomalies, and the modes of the anomalies can be known, for example, the local anomalies may be caused by traffic jam or traffic accidents, and the global anomalies may be global fraud behaviors of taxi drivers.

Description

technical field [0001] The invention belongs to the technical field of track detection and provides a method for detecting abnormal tracks based on fuzzy theory. Background technique [0002] Vehicles embedded with GPS devices collect a large amount of vehicle trajectory data, making it possible to analyze user behavior through user trajectory data. For these reasons, research on user behavior analysis in the field of intelligent transportation systems becomes very important. To meet the needs of a wide range of applications, researchers have proposed many techniques and methods to mine valuable information from large-scale trajectory data, such as location-based services, transportation management, traffic flow prediction, and urban planning. In this case, trajectory data mining has increasingly become an important research topic, mainly including trajectory classification, trajectory clustering, trajectory prediction, trajectory pattern mining and outlier detection. In t...

AI Technical Summary

Problems solved by technology

[0003] Car rental fraud by TAXI services in modern cities has resulted in many complaints and can lead to a bad name for the taxi service, such as taxi drivers charging higher rates for time, taking passengers not on the usual routes, choosing to take detours, detours The distance is usually longer than the normal path between a pair of source and destination, but the existing technology can not detect the detour fraud of the taxi, so it is impossible to identify or track the detour fraud of the outgoing car, and thus cannot detect it. Supervision is not conducive to the improvement of the quality of taxi services

Images