An effective conflict management method based on belief similarity measure and entropy for multi-sensor data fusion

Multi-sensor data fusion has received substantial attention thanks to its ability to integrate information from distinct sources efficiently. Nevertheless, the information collected from multi-sensors may be uncertain and imprecise, even conflicting in real applications. As a distinguished theory to handle uncertain and imprecise information, belief functions theory (BFT) is prevalent in the various fields of multi-sensor data fusion. Unfortunately, counter-intuitive behaviors may generate once facing highly conflicting pieces of evidence. To deal with the above-mentioned issue, in this paper, we study a novel belief Sørensen coefficient (\(\mathcal {BSC}\)) to measure the conflict between the pieces of evidence based on BFT. On the top of \(\mathcal {BSC}\), we propose a new belief conflict coefficient, and prove some important properties, namely, non-negativity, symmetry, non-degeneracy, bounded, extreme consistency and insensitivity to refinement. In parallel, some numerical examples are employed to demonstrate the superiority of the belief conflict coefficient in quantifying the degree of conflict between the pieces of evidence. Finally, we design a new multi-sensor data fusion method based on the proposed \(\mathcal {BSC}\) and the improved belief entropy, and verify the effectiveness and practicability of the proposed method with respect to other methods through several application cases.