Data-mining-based Identification of Behavior of Drag Suction Dredger

A method of unsupervised identification of the behavior of drag suction dredger, which features frequently changing operation mode and closely distributed sailing trajectory. Since AIS information is usually not available, an unsupervised identification method is adopted. The identification algorithm is developed based on DBSCAN. The core point of clustering is defined according to the average speed instead of density and threshold range is set for different features. The multi-mode synchronized clustering achieves highly efficient identification. The operation location, dredged area, time spent and other performance indicators are further analyzed based on the identification output.