A Comparative Analysis of Evidential Reasoning and Cumulative Scoring Algorithms in the Context of a Combat Identification Application

Within the discipline of Combat ID, as related to high performance fighters, a problem of significant proportion that arises is how to determine quickly and accurately a target’s ID. Various algorithms, many of which are proprietary, that address this problem have been developed over the last decade. But their robustness in handling a new and unforeseen target threat environment leaves much to be desired. In this paper we explore two candidate classes of algorithms for achieving this objective, namely, an evidential reasoning and a cumulative scoring technique. We present the results via a comparative analysis where we look at the problem not only from a timeliness and accuracy of identification perspective, but also from the point of view of computational throughput. The analysis is based on simulation results using the sensor fusion system of a high performance fighter program. This system accomplishes the fusion of attribute data from a diverse set of sensors in a real time, computationally constrained processing environment. We will show some of the performance advantages evidential reasoning exhibits over a cumulative scoring approach. This is demonstrated on a typical scenario that the fighter in question has to show performance against. Furthermore, we will discuss a methodology for using a by-product of the evidential reasoning algorithm as a score to help in the data association task of assigning sensor reports to system tracks.