Multi-hypothesis structures and taxonomies for combat identification fusion

Abstract

One of the greatest difficulties in developing a fusion process is determining the type, quantity, and quality of the information provided. Even when this is accomplished, the utility (relationship) of the information is often difficult to establish. For the problem of combat identification (combat ID or CID) this is especially taxing. Often numerous sources provide information, but relationship guidelines are not well developed, or are ambiguous or inconsistent. This deficiency leads to poorly constructed fusion architectures and methodologies because information is either ignored or improperly combined in the fusion process. Using the Joint Directors of Laboratories (JDL) information fusion model as a guide, this paper address the movement of attribute information across multiple hypothesis classes (with some examination of kinematic information) as it relates to developing the identification of different objects, and how it can be combined both within and between JDL fusion levels. The result of this analysis leads to an information architecture that is naturally adaptive to information regardless of quality, level, or specificity.