Event processing has been established as a generic computational paradigm in a wide range of applications. Events report on state changes of a system and its environment, thereby enabling reactive and proactive computing. At the core of event processing systems is an event recognition mechanism. It is the ability of a system to detect events that are considered relevant for processing and, as such, is the basis of situation awareness.
The goal of this project is the development of generic event recognition technology that may be utilized for maritime surveillance. To deal with the lack of veracity, probabilistic reasoning techniques will be employed. To address data volume, velocity and distribution, novel algorithms for distributed reasoning under uncertainty will be developed. Moreover, to allow for proactivity, event forecasting techniques will be proposed.
The generic event recognition and forecasting technology will be evaluated using real data streams from the maritime domain.