An optimal fusion processor for multiple radar surveillance

This paper presents an optimal fusion criterion for merging statistical decisions of multiple CFAR sensors. The fusion rule depends on the sensor specifications and the probability model imposed on the observables. Optimality is defined in terms of minimizing the total expected cost incurred by the fusion center´s decision. By assigning appropriate costs to each decision, a minimum error probability fusion rule is derived. Numerical studies on a system employing two localized CFAR sensors demonstrates the global performance of the fusion rule