1: \begin{abstract}

2: In this paper, we develop a consensus algorithm for distributed computation of the \ac{rcm} on Lie Groups. The algorithm is built upon a distributed optimization reformulation that allows developing an intrinsic, distributed (without relying on a consensus subroutine), and a computationally efficient protocol for the \ac{rcm} computation. The novel idea for developing this fast distributed algorithm is to utilize a Riemannian version of distributed gradient flow combined with a gradient tracking technique. We first guarantee that, under certain conditions, the limit point of our algorithm is the \ac{rcm} point of interest. We then provide a proof of global convergence in the Euclidean setting, that can be viewed as a ``geometric'' dynamic consensus that converges to the average from arbitrary initial points. Finally, we proceed to showcase the superior convergence properties of the proposed approach as compared with other classes of consensus optimization-based algorithms for the \ac{rcm} computation.

3: \end{abstract}

4: