On a generic fusion model in situation assessment processes dealing with disparate sources

Situation assessment process like identifying or localizing a target consists in recognizing one situation out of a set of possibilities. In operational context, multisensor analysis of the situation requires to take into account the uncertainties induced by unfavorable conditions of learning or models, erroneous, incomplete or imprecise measurements… . Beside sensor measurements, the fusion process is intended to manage any exogenous contextual information able to reflect learning model validity and hence overcome the individual insufficiencies of each sensor. A generic modeling of this type of information in the form of mass sets of theory of evidence, has formerly been proposed by Appriou. A closer attention has been firstly accorded to the most common case where the data originate from statistical learning. This paper is concerned with a multisensor analysis dealing with disparate data expressed in different theoretical formalisms according to their specificities in terms of uncertainty and/or imprecision. The proposed developments are focusing on the interpretation and modeling of such different types of observation and prior knowledge in the form of appropriate mass sets in such a way they can be merged within the evidence theory framework, leading to decision criteria able to handle the disparity of information