Active Information Fusion	Efficient Sensor Selection for Active Information Fusion In active information fusion, determining the most informative and cost-effective sensors requires an evaluation of all possible sensor combinations, which is computationally intractable. We present a methodology to actively select a sensor subset with the best tradeoff between information gain and sensor cost by exploiting the synergy among sensors. Our approach includes two aspects: a method for efficient mutual information computation, and a graph-theoretic approach to reduce search space. The approach can reduce the time complexity significantly in searching for a near optimal sensor subset. In the proposed sensor selection algorithm, the first step is to construct a synergy graph for all the sensors. The synergy graph is G=(S, E), where S are nodes, representing the set of available sensors, and E are edges, representing the set of pairwise synergetic links weighted by synergy coefficients between the corresponding pairwise sensors. If all sensors in a subset on G are serially linked, this subset of sensors is referred to as a sensor synergy chain. Initially, the synergy graph is a completely connected network. Rules are then developed to prune the synergy graph such that unpromising sensor combinations are removed from the graph. Search is then conducted in the pruned graph to identify the optimal sensor combination. More ...
Introduction
Summary
Demos
Publications
BN Resources