Trust and independence aware decision fusion in distributed networks

In distributed network environments, decisions must often be made based on incomplete or uncertain evidence whose sources may be dependent. Properly fusing potentially unreliable and dependent information from multiple sources is critical to effective decision making. The Transferable Belief Model (TBM), an extension of Dempster-Shafer Theory (DST), is a well known information fusion framework that can cope with conflicting evidences. However, neither DST nor TBM deals with misbehaving data sources and dependence of fusion data, which are often observed in dynamic multi-hop network environments. In this work, we propose a decision fusion framework that considers multi-dimensional trust and independence of information using a provenance technique, to enhance fusion reliability. We consider three information trust dimensions: correctness, completeness, and timeliness. Our simulation results show that the proposed framework yields a higher correct decision ratio, compared with the baseline counterparts.