Handling Data Uncertainty Using Multi Entity Bayesian Networks in Data Fusion

The goal of high-level data fusion is the perception of the elements in the environment, the comprehension of their meaning, and the projection of future status before decision making. There are many kinds of uncertainty involved in data fusion systems. Consequently, it is important to have consistent and principled techniques to manage uncertainty. Bayesian Networks (BNs) have been successfully applied to create probabilistic representations of uncertain knowledge in diverse fields. Moreover, Multi Entity Bayesian Networks (MEBNs) extend the propositional expressiveness of BNs to achieve the representational power of first-order logic. In this paper the use of MEBN for handling uncertainty in high-level data fusion is examined, and further explanations are mentioned through a case study on fusing reports from various sources to identify type of military vehicle.