Bayesian network model for data incest in a distributed sensor network
McLaughlin S., Krishnamurthy V., Evans RJ.
This paper addresses the problem of linear estimation in a distributed sensor network. A fundamental issue in many-to-many sensor configurations is that of data incest, arising from the inadvertent (or ignored) multiple use of identical information. Extending recent work by the authors that considers the transmission of a global estimate, this paper casts the decentralized estimation problem in a Bayesian Network framework. A fusion strategy is presented in this framework, to address the problem of estimating a dynamic target in a network where variable delays exist between nodes.