Detecting anomalous behavior in sea traffic: A study of analytical strategies and their implications for surveillance systems

Surveillance operators normally analyze vast amounts of sensor data in order to find conflict situations, and threatening or unusual activities while allowing the continuous flow of goods and people. Semi-automatic support may reduce the time needed for the detection of such situations, generating early warnings that can prevent accidents or provide time to prepare countermeasures. In order to provide adequate cognitive support for operators and guide the design of more efficient surveillance systems, this paper investigates the human analytical reasoning process of detecting anomalous behavior through a case study, the surveillance of sea areas. The analysis of data gathered during interviews and participant observations at three maritime control centers and the inspection of video recordings of real incidents lead to a characterization of operators' analytical processes. We suggest how to support these processes using data mining and visualization, and we derive recommendations for designers and developers of future maritime control systems.