Acquiring environmental information yields better anticipated search performance
Harun Yetkin, Collin C. Lutz, Daniel J. Stilwell · 2016
This paper addresses the problem of finding an unknown number of stationary objects distributed in a bounded search domain within bounded time. We assume the search agent is equipped with a search sensor and an environmental characterization sensor, and that these sensors operate simultaneously. Our cost function accounts for false positives, false negatives and environmental uncertainty. We consider that environmental conditions affect the performance of the search sensor, and uncertainty in the environment causes the anticipated search performance to be different than the actual search performance. Our approach aims to minimize the deviation from actual search performance while maximizing the probability that our estimates on the number of objects will be correct.