Fusion of Imperfect Information in the Unified Framework of Random Sets Theory: Application to Target Identification.

摘要

This is a study of the applicability of random sets theory to target identification problems as a technique for fusion of imperfect information. For target identification, several sources of information (radar, ESM - Electronic Support Measures, SAR - Synthetic Aperture Radar, IR images) are available. Since the information provided is always imperfect and several kinds of imperfection may be encountered (imprecision, uncertainty, incompleteness, vagueness, etc.), several theories were developed to assist probability theory (long the only tool available to deal with uncertainty) in data fusion problems. In recent decades fuzzy sets theory was developed to deal with vague information, possibility theory was developed to deal with incomplete information, evidence theory was developed to deal with imprecise and uncertain information, and rough sets theory was developed to deal with vague and uncertain information. These theories have several points in common; here we study random sets theory, which is a unifying framework for all the aforementioned theories. In two simple test scenarios, we demonstrate the effectiveness of this unifying framework for representing and fusing imperfect information in the target identification application.