DSmH Evidential Network for Target Identification ⋆

This paper proposes a model of evidential network based on Hybrid Dezert-Smarandache theory (DSmH) to improve target identification of multi-sensors. In the classification simulation, we compared the results obtained at the Target Type node and Foe-Ally node in evidential network by using DempsterShafer theory (DS) and using DSmH. The comparisons show that, when we use DSmH in the evidential network, we can assign more Basic Belief Assignments (BBA) to the focal element the target belongs to. Experiments confirm that the model of evidential network using DSmH is better than the one using DS.