Analysis of CBRN sensor fusion methods

Many well-defined metrics have been developed for data fusion problems in the target tracking domain. However, much less is known about the proper evaluation of Chemical, Biological, Radiological, and Nuclear (CBRN) data fusion methods. In this paper we begin to address the issue of evaluating the performance of CBRN fusion algorithms. As a demonstration of the applicability of modified target tracking metrics to CBRN problems we apply such techniques to an illustrative data fusion method. Of the metrics we have studied one stands out among the rest, namely covariance consistency, and we pay special attention to its proper application to CBRN scenarios. Covariance consistency is of particular importance since it allows us to evaluate how accurately an algorithm is reporting the uncertainty of its estimates, and we would contend that uncertainty measures are of paramount importance in CBRN scenarios.