Object identification using T-conorm/norm fusion rule

This small chapter presents an approach providing fast reduction of total ignorance in the process of target identification. It utilizes the recently defined fusion rule based on fuzzy T-conorm/Tnorm operators, as well as all the available information from the adjoint sensor and additional information obtained from the a priori defined objective and subjective considerations, concerning relationships between the attribute components at different levels of abstraction. The approach performance is estimated on the base of the pignistic probabilities according to the nature of the objects considered here. The method shows better efficiency in comparison to the pure Dempster-Shafer theory based approach. It also allows to avoid the application of the Bayesian principle of indifference and improves the separation power of the decision process. 1 This work is partially supported by MONT grant MI-1506/05.