Advance in Multiply Sectioned Bayesian Networks: Sensor Network Practitioners´ Perspective

Multiply sectioned Bayesian networks provide a probabilistic framework for reasoning about uncertain domains in cooperative multiagent systems. Several advances have been made in recent years on modeling, compilation and inference under the framework. This paper links these advances together through a case study and presents them from the perspective of practitioners in intelligent sensor networks. We demonstrate how the framework can be applied to multisensor fusion and how intelligent sensor agents developed by independent vendors can be integrated into a coherent sensor fusion system.