Active-testing surveillance for multiple target detection with composite hypotheses

摘要

We consider a multiple target detection problem in which there are N target sites, or cells, which can be energized independently, while the total energy available is finite. In the specific model used here, the data are conditionally Poisson, with the mean being the product of the transmitted energy and the gamma distributed reflectivity. A Bayesian sequential analysis procedure for target detection is proposed and strategies for partitioning the available energy at each iteration are compared.