Maritime Observation Systems (AIS, radar, …) generating a lot of data, which have a great potential to gain new insights and a better understanding of the maritime traffic. Identifying anomalies (e.g. unplanned maneuvers) for test case generation is an important topic in the development process of highly automated systems. For anomaly detection it is necessary to identify common behavior, to separate normal from anomal maneuvering in navigational waterways. By observing vessel tracks, the behavior can be derived from the driven maneuvers of a vessel. Extracting maneuvers out of historical observation data (e.g. AIS) is an important task. In the following work the authors propose the Maneuver Net generation process for deriving the sailing behavior and how Maneuver Nets can be used for unplanned maneuver detection.
Paper
Full text
Statistical Maneuver Net Generation for Anomaly Detection in Navigational Waterways
Semantic Scholar · Engineering · 2019
Abstract
Maritime Observation Systems (AIS, radar, …) generating a lot of data, which have a great potential to gain new insights and a better understanding of the maritime traffic. Identifying anomalies (e.g. unplanned maneuvers) for test case generation is an important topic in the development process of highly automated systems. For anomaly detection it is necessary to identify common behavior, to separate normal from anomal maneuvering in navigational waterways. By observing vessel tracks, the behavior can be derived from the driven maneuvers of a vessel. Extracting maneuvers out of historical observation data (e.g. AIS) is an important task. In the following work the authors propose the Maneuver Net generation process for deriving the sailing behavior and how Maneuver Nets can be used for unplanned maneuver detection.