Research Article
A Collaborative Anomaly Detection Approach of Marine Vessel Trajectory (Short Paper)
@INPROCEEDINGS{10.1007/978-3-030-30146-0_19, author={Zejun Huang and Jian Wan and Jie Huang and Gangyong Jia and Wei Zhang}, title={A Collaborative Anomaly Detection Approach of Marine Vessel Trajectory (Short Paper)}, proceedings={Collaborative Computing: Networking, Applications and Worksharing. 15th EAI International Conference, CollaborateCom 2019, London, UK, August 19-22, 2019, Proceedings}, proceedings_a={COLLABORATECOM}, year={2019}, month={8}, keywords={Vessel trajectory Collaborative anomaly detection Multi-factor Marine}, doi={10.1007/978-3-030-30146-0_19} }
- Zejun Huang
Jian Wan
Jie Huang
Gangyong Jia
Wei Zhang
Year: 2019
A Collaborative Anomaly Detection Approach of Marine Vessel Trajectory (Short Paper)
COLLABORATECOM
Springer
DOI: 10.1007/978-3-030-30146-0_19
Abstract
Trajectory anomaly detection plays a very important role in navigation safety. Most trajectory anomaly detection methods mainly detect the spatial information of the vessel’s trajectory. These methods neglect a vessel’s dynamic behavior characteristics, such as course, speed, and acceleration. In this paper, a vessel trajectory multi-factor collaborative anomaly detection (VT-MCAD) approach is proposed to realize the anomaly detection of vessels at sea by studying the trajectory characteristics of vessels. Firstly, the trajectory behavior of historical vessels is identified, and the trajectory characteristics, such as course, speed, and acceleration, are extracted for different trajectory behaviors. Then, the current trajectory behavior is identified when the trajectory anomaly is detected. Based on the TRAjectory Outlier Detection (TRAOD) method, the corresponding trajectory feature model components, including instantaneous angle acceleration, average angle acceleration, instantaneous velocity, and the average velocity and acceleration, are used to detect the anomaly trajectory, and trajectory’s suspicious degree of each component in VT-MCAD are obtained. Finally, the suspicious degree of each component is combined to calculate the final suspicious degree. VT-MCAD can change the weight of components according to the detection effectiveness of different components and avoid excessive dependence on one component, which results in better robustness and reliability. The experimental results based on real-world vessel data showed that VT-MCAD could effectively capture anomaly trajectories.