Detecting Anomalous Trajectories and Behavior Patterns Using Hierarchical Clustering from Taxi GPS Data

Abstract

Anomalous taxi trajectories are those chosen by a small number of drivers that are different from the regular choices of other drivers. These anomalous driving trajectories provide us an opportunity to extract driver or passenger behaviors and monitor adverse urban traffic events. Because various trajectory clustering methods have previously proven to be an effective means to analyze similarities and anomalies within taxi GPS trajectory data, we focus on the problem of detecting anomalous taxi trajectories, and we develop our trajectory clustering method based on the edit distance and hierarchical clustering. To achieve this objective, first, we obtain all the taxi trajectories crossing the same source-destination pairs from taxi trajectories and take these trajectories as clustering objects. Second, an edit distance algorithm is modified to measure the similarity of the trajectories. Then, we distinguish regular trajectories and anomalous trajectories by applying adaptive hierarchical clustering based on an optimal number of clusters. Moreover, we further analyze these anomalous trajectories and discover four anomalous behavior patterns to speculate on the cause of an anomaly based on statistical indicators of time and length. The experimental results show that the proposed method can effectively detect anomalous trajectories and can be used to infer clearly fraudulent driving routes and the occurrence of adverse traffic events.

Detecting Anomalous Trajectories Using the Dempster-Shafer Evidence Theory Considering Trajectory Features from Taxi GNSS Data

“…In the field of traffic safety, traffic events are detected by using video detection data in urban traffic automatically, and the relationship between events and conflict is analyzed [7][8][9]. For GNSS trajectory data, they have two areas of application: one is to discover anomalous traffic [10,11], and the other is to infer anomalous trajectories [12][13][14][15]. Our research focuses on the detection of anomalous trajectories.…”

“…In recent years, anomalous trajectory detection has included the distance-based method [16,17], the clustering-based method [12,18,19], the classification-based method [20,21], and the grid-based method [13][14][15]. The distance-based method divides the whole trajectory into sub-trajectories that include certain trajectory points and calculates the Hausdorff distance between sub-trajectories.…”

“…The similarity of the trajectory is mostly described by considering the distance between the trajectory points or the trajectory segments. Commonly used distances include, for example, the Euclidean distance, dynamic time warping (DTW), and edit distance (ED) [12]. The trajectories are unevenly distributed, and it is difficult to exhibit aggregation.…”

“…The trajectories are unevenly distributed, and it is difficult to exhibit aggregation. The research data are almost all trajectories from same source-destination pairs [12]. The classification-based method builds a classification model based on trajectory features and requires more sample data for learning and training.…”

A Simple Line Clustering Method for Spatial Analysis with Origin-Destination Data and Its Application to Bike-Sharing Movement Data

“…These components were adapted from similar measures that were used for pattern recognition [15]. Some researchers have adopted this definition and made some improvements [12,[16][17][18][19]. However, the physical meaning of this definition is not very clear, especially in the weight setting section.…”