The quantization error in the average consensus problem

This work studies how uniform quantization in communications affects the linear diffusion algorithm for average consensus problem. Starting from the well-known linear diffusion algorithm, we analyze a simple and effective adaptation which is able to preserve the average of states and to drive the system near to the consensus value. The error is estimated by a worst case analysis, which suggests that it can increase as the number of agents goes to infinity. The algorithm is also compared with the existing literature.