Reducing impact of conflicting data in DDFS by using second order knowledge

Fusing estimation information in a Distributed Data Fusion System (DDFS) is a challenging problem. One of the main issues is how to detect and handle conflicting data coming from multiple sources. In fact, a key of success of a Data Fusion System is the ability to detect wrong information. In this paper, we propose the inclusion of reliability assessment of information sources in the fusion process. The evaluated reliability imposes constraints on the use of information data. We applied our proposal in the challenging scenario of Multi-Agent Multi-Object Tracking. © Springer-Verlag Berlin Heidelberg 2010.