Discovery of Chasing Patterns in Trajectory Data

Abstract

Trajectory data contains abundant temporal and spatial information of moving objects. By calculating and analyzing trajectory data, behavior patterns of mobile objects can be found. Among these behavior patterns the discovery of chasing patterns is interesting which will provide effective services for intelligent urban traffic management, criminal investigation, environmental monitoring and other application fields. However, existing chasing pattern discovery methods only compare the local similarity of trajectories, and do not consider the potential significance of stay points or change in speed. This article proposes a new algorithm for chasing pattern discovery based on the stay point detection technology. An optimized stay point detection algorithm is also designed to improve the distance calculation method. The algorithms were evaluated with real and simulation trajectory data and achieved very good results which have higher accuracy and recall rate compared with the existing chasing pattern discovery algorithms.