Efficient employment of adaptive sensor

Abstract

We consider a sensor that is subject to false positive and false negative errors. The sensor searches for stationary threat objects such as ballistic missile launchers or improvised explosive devices. The objects are located in a certain area of interest, which is divided into areacells. The areacells are defined such that each of them may contain, at most, one threat object. The task of the sensor is to determine if an areacell contains a threat object, and the objective of the searcher is to maximize the number of correctly determined areacells. Since definitive identification of a threat object, and subsequent handling of that threat, are done by a limited number of available ground combat units, the correct determination of an areacell is crucial for better allocating and directing these scarce resources. We develop an algorithm, rooted in large deviations theory and stochastic approximation theory, that leads to the optimal search effort. The computed allocation maximizes the expected number correctly determined areacells as the number of available looks for searching becomes large.