Decentralized Data Fusion and Control in Active Sensor Networks

The paper presents an algorithm for Bayesian decentralized data fusion (BDDF) and its extension to informationtheoretic control. The algorithm is stated for a feature represented by a general probability density function. Several specific representations are then considered – Gaussian, discrete, Certainty Grid, and hybrid. Well known algorithms for these representations are shown to fit the general BDDF pattern. Stating the algorithms in Bayesian terms has a practical advantage of allowing a generic software implementation. It is also hoped that a clear general formulation will stimulate extensions to efficient non-parametric representations of arbitrary distributions. The algorithms are described in the context of the Active Sensor Network architecture – a modular framework for decentralized cooperative data fusion and control. The approach is illustrated with the results of two deployment scenarios with an indoor sensor network.