Distributed estimation with data association: Is the nearest neighbor the most informative?

In distributed multi-sensor estimation/tracking the problem of measurement fusion arises. In large sensor networks (SN), each sensor is constrained by bandwidth to communicate only one of its observations to a fusion center (FC) for a global estimate. We study the problem of distributed estimation with data association, where the FC ldquooptimallyrdquo combines the ldquobestrdquo measurements from the sensors, instead of suboptimally combining the local estimates. Using order statistics, we show that, surprisingly, the nearest neighbor (NN) is not always the most informative measurement. Simulations corroborate our analysis.