A scalable wireless communication architecture for average consensus

This paper introduces a multiple-access coding technique that is tailored to solve average consensus problems efficiently in wireless networks. We propose a novel data driven architecture which grants channel access to nodes based on their local data values. We analyze the performance of the scheme in the presence of quantization errors and noise. We show that our scheme is unbiased with respect to quantized consensus algorithms, it achieves good MSE performance, and it can be configured to provide a speedup in the convergence rate. The amount of speedup achieved is a function of |Q_k| which indicates the number of quantization bins used to represent the state variables exchanged during the computation.