Evidential networks for decision support in surveillance systems

The chapter is devoted to the development of an automatic knowledge-based information fusion system that supports the decision making in a reliable, timely and consistent manner. The context is military and homeland security domain, particularly challenging due to the inherently incomplete, uncertain and imprecise evidence and domain knowledge. The mathematical tool used in the chapter is the valuation algebra in the context of the theory of evidence. The motivation is the expressive power of belief functions which can represent both classical probability functions and possibility/necessity functions. The resulting network for reasoning under uncertainty is referred to as the evidential network. The theoretical concepts are demonstrated by two examples: decision support systems for target identification and threat assessment.