1: \begin{abstract}

2: Belief propagation (BP) is a powerful tool to solve distributed inference

3: problems, though it is limited by short cycles in the corresponding

4: factor graph. Such cycles may lead to incorrect solutions or oscillatory

5: behavior. Only for certain types of problems are convergence properties

6: understood. We extend this knowledge by investigating the use of reweighted

7: BP for distributed likelihood fusion problems, which are characterized

8: by equality constraints along possibly short cycles. Through a linear

9: formulation of BP, we are able to analytically derive convergence

10: conditions for certain types of graphs and optimize the convergence

11: speed. We compare with standard belief consensus and observe significantly

12: faster convergence. 

13: \end{abstract}

14: