Information quality evaluation in fusion systems

Advances in new information and communication technology have considerably increased the quantity of available data and permitted the implementation of more and more complex processing techniques. As a result of these changes it appeared the need of evaluating the information fusion system quality and to propose to the user the information accompanied with confidence coefficients. The information fusion systems used nowadays are complex systems that are difficult to be evaluated. To overcome this problem we propose a new quality evaluation methodology based on the decomposition of the information fusion system in its elementary modules that allows the evaluation of the quality at two levels. The first one (global) that describes the entire information fusion system and the second one (local) for each elementary module. As data and information change over time, this decomposition allows to directly evaluate the global quality of the information fusion system using the local quality evaluation.