Distributed Consensus With Quantized Data via Sequence Averaging

The problem of distributed average consensus with quantized data is considered in this correspondence. Conventional consensus algorithms suffer from divergence when quantization errors are present. To address this issue, we introduce a modified quantization-based consensus protocol and exploit the temporal information collected from the iterative process, based on which we develop an efficient consensus algorithm. The proposed consensus algorithm is proved to converge to the true mean, i.e., the average of the initial state, in a mean square sense. It also presents an advantage of speeding up the convergence over the algorithm [P. Frasca, R. Carli, F. Fagnani, and S. Zampieri, ¿Average Consensus on Networks With Quantized Communication,¿ Int. J. Robust Non-Linear Control, 2008, to be published] without exploitation of temporal information. Numerical results are presented to illustrate the effectiveness of the proposed algorithm.