Decision support for wide area search in a radiological threat scenario: for intelligent reachback using complex DSS architectures

Abstract

In the stressful scenario of a terrorist threat that involves a radiological dispersion device, an effective search strategy must be prepared. This problem is highly complex, especially when multiple platforms are deployed. Besides the automated suggestions for an effective search strategy, a remote advice and assist cell can benefit from visual aid in the form of a heat map overlay on a geographical map, which evolves according to both the expected target movements and the observations made by the sensors. Such decision support software can be equipped with optimization algorithms that provide search strategies. However, the algorithms are only as good as the accuracy of the parameter configuration, such as the estimation of the target motion and the search effectiveness of the sensor. In this paper, we focus on the estimation of the latter parameter for two sensors and two types of targets. The first step in this process consists of field experiments to approximate the target dependent sensor ranges. We then describe a fictive but realistic scenario in which a suspect carrying radiological material must be detected to prevent a terrorist attack in the center of a big city. We show how to estimate the parameters for the search effectiveness in this scenario. Finally, in a proof of concept, we compare two algorithms for optimization of a search strategy for multiple searchers. Here, the algorithms are configured using the mentioned parameter estimations. Computational experiments show that the search effectiveness significantly influences the probability of detection throughout the search.