An Asynchronous Trajectory Matching Method Based on Piecewise Space-Time Constraints

Abstract

Automatic Identification System (AIS) is an automatic tracking system based on reports provided by AIS transponders, which can report the static and dynamic information of the ship, such as name, call sign, position and speed etc. However, AIS can be manually turned off, making it impossible to find the trajectory of the ship when AIS is turned off, which results in difficulties in maritime supervision. This article uses shore-based radar trajectory data to present a new method of asynchronous trajectory matching based on piecewise space-time constraints (PTSCTM), which is used to reconstruct AIS trajectory and discover the abnormal behavior of ships. The Ramer-Douglas-Peucker (RDP) algorithm is used to extract feature points of the trajectory, and the trajectory is segmented according to feature points. The method uses space-time constraints to determine candidate trajectory matching points in each trajectory segment and calculates the optimal matching point from the candidate set through the space and time distance. The method asynchronously calculates the similarity of trajectory segments from multi-source trajectories and calculates the trajectory similarity to achieve multi-source trajectory matching. The measured data show that the model can effectively measure the similarity of multi-source trajectories to find the optimal radar trajectory as a matching result when AIS is turned off, and the calculation accuracy and efficiency of the model is higher than the existing methods. According to the asynchronous trajectory matching method, the monitoring method of staying ships in the Qingdao lancelet reserve area is proposed. The result shows that many ships stayed in the area through radar signals when AIS was turned off. The result can provide auxiliary support for maritime supervisory and legal departments, which helps to strengthen maritime ship supervision, detect maritime illegal crimes and protect the marine environment.