A New Local Measure of Disagreement between Belief Functions – Application to Localization Arnaud Roquel, Sylvie Le H´ egarat-Mascle, Isabelle Bloch, and Bastien Vincke Abstract. In the theory of belief functions, the disagreement between sources is often measured in terms of conﬂict or dissimilarity. These measures are global to the sources, and provide few information about the origin of the disagreement. We propose in this paper a “ﬁner” measure based on the decomposition of the global measure of conﬂict (or distance). It allows focusing the measure on some hypothe- ses of interest (namely the ones likely to be chosen after fusion). We apply the pro- posed so called “local” measures of conﬂict and distance to the choice of sources for vehicle localization. We show that considering sources agreement/disagreement outperforms blind fusion. 1 Introduction Multi-sensor systems are used in many applications such as classiﬁcation, image processing, change detection, object trajectory localization. Usually the informa- tion provided by each sensor is prone to imperfections, such as imprecision and uncertainty, and fusion procedures aim at making better decisions by combining multi-sensor information. Belief Functions (BF) are suitable for modeling impreci- sion and uncertainty, and handle belief on the power set of the frame of discernment (set of hypotheses). A disagreement between sources makes the system unstable and can impact the decision. Many techniques have been developed to measure the disagreement between sources. A review can be found in [4] or [5]. One method con- sists in observing the so-called “Demspter’s conﬂict” [10] resulting from the con- junctive combination of the basic belief functions. However, the non-idempotence Arnaud Roquel · Sylvie Le H´ egarat-Mascle · Bastien Vincke Universit´ e Paris Sud, IEF, Orsay, France e-mail: first-name.last-name@u-psud Isabelle Bloch T´ el´ ecom ParisTech, CNRS LTCI, Paris, France e-mail: isabelle.bloch@telecom-paristech.fr T. Denœux & M.-H. Masson (Eds.): Belief Functions: Theory & Appl., AISC 164, pp. 335–342. springerlink.com c  Springer-Verlag Berlin Heidelberg 2012