Situation assessment based on spatially ambiguous multisensor measurements

In most of the requirements met in situation assessment, sensor analysis ha s to be able to recognize in adverse conditions one hypothesis out of a set of hypotheses concerning, for instance, either the localization, the ident ity, or the number of targets. Implementing multiple complementary sensors is then of major interest, especially if they have orthogonal spatial resol utions. Nevertheless, when dealing with multiple targets in an area of inte rest this leads to ambiguities in data association, due to ghosts (erroneou s matching), hidden targets, non-detections, and false alarms, and requires a suitable matching process. Furthermore, measurements may have a doubtful origin, and prior knowledge is understood to be often poorly defined, unce rtain, or incomplete. So, the present synthesis proposes a generic modeling of this type of information in the framework of the theory of evidence, wi th closer attention being paid to the different ways data are processed in common cases. The decision process required by any assessment function for the identification of the most likely hypotheses is then discussed, and a g lobal approach of the detection, matching, counting, localization, and clas sification functions is proposed that aims at exploiting all useful knowled ge. The complementary contribution of the different sources of information involved, as well as the suitability of their processing, can be emphasized on a didactic example. (C) 2001 John Wiley & Sons, Inc.