An increasing number of ships are equipped with vessel monitoring systems logging ship data during normal operation. We developed a framework for multivariate time series data mining to extract the information of vessel behavior from an in-service dataset. The approach is established on unsupervised data clustering using Self-Organizing Map (SOM), K-means, and k-Nearest Neighbors Search (K-NNS) for searching specific maneuvers. The results are based on ship monitoring data of NTNU’s research vessel, Gunnerus. It is shown that this approach is effective to detect prior unknown ship states with acceptable accuracy. The framework proposed and the results of this work can be of interest to those involved in ship administration, marine traffic flow engineering, ship maneuvering studies and assessment of ship design.
A Data Mining Approach to Identify Maneuvers in Ship In-Service Measurements
- Views Icon Views
- Share Icon Share
- Search Site
Abbasi Hoseini, A, & Steen, S. "A Data Mining Approach to Identify Maneuvers in Ship In-Service Measurements." Proceedings of the ASME 2016 35th International Conference on Ocean, Offshore and Arctic Engineering. Volume 7: Ocean Engineering. Busan, South Korea. June 19–24, 2016. V007T06A088. ASME. https://doi.org/10.1115/OMAE2016-54180
Download citation file: