Image: Enno Kapitza for DLR
In order to achieve the above goal, the team uses data and information from the automatic identification system (AIS), from the navigational sensors of the vessel (working group multi sensor systems) and from the maritime radar. All systems or sensors have properties which are supporting the nautical personal during the task of traffic situation assessment. Each of these systems can be operated in an autarkic mode on the bridge of the vessel but with restrictions (for example AIS data can be corrupt or wrong or the radar screen might be dominated by sea clutter). These restrictions of single sensor systems explain the need for sensor fusion technologies, which are able to mitigate those limitations.
Before it is possible to fuse systems or sensors it is of crucial importance to know the error behavior of each single system or sensor. Furthermore the rapid growth of computational power as well as the progress in the field of network based systems allows the use of numerical methods and algorithms in real time, which was not feasible just a few years ago.
Insofar, the current research activities of the team is focused on answering the question of, which object tracking procedure is able to reliably track maritime targets in extreme conditions (multiple targets, cluttering). In order to improve the performance of the developed algorithms and processors we use a traffic situation simulation system, which is able to simulate a variety of different scenarios in a reproducible manner. We are continuously working on the improvement of our algorithms in the research harbor Rostock running AIS plausibility monitor focusing on the extension of the AIS for collision avoidance. Our research as well focuses on (multiple) extended target tracking in maritime environments. We also provide a couple of marine radar datasets to the research community. All these activities eventually aim at the development of robust fusion and tracking algorithms for reliable assessment of the traffic situation.