Can we trust subjective logic for information fusion?

J. Dezert (1) , A. Tchamova (2) , D. Han (3) , J.M. Tacnet (4)

1 ONERA - The French Aerospace Lab [Palaiseau]
2 IICT BAS - Institute for Information and Communication Technologies
3 Xjtu - Xi'an Jiaotong University
4 UR ETGR (ETNA) - Erosion torrentielle neige et avalanches

In this paper, we provide a deep examination of the main bases of Subjective Logic (SL) and reveal serious problems with them. A new interesting alternative way for building a normal coarsened basic belief assignment from a refined one is also proposed. The defects in the SL fusion rule and the problems in the link between opinion and Beta probability density functions are also analyzed. Some numerical examples and related analyses are provided to justify our viewpoints.

Mots clés: SUBJECTIVE LOGIC; BELIEF FUNCTIONS; INFORMATION FUSION

Domaines: Logique en informatique [cs.LO]

J. Dezert, A. Tchamova, D. Han, J.M. Tacnet. Can we trust subjective logic for information fusion?. Fusion 2014, Jul 2014, SALAMANCA, Spain. ⟨hal-01070401⟩