Signal-centric predictive polling for medium access control in M2M communications networks

In this paper, we propose signal-centric polling-based medium access control algorithms for machine-to-machine communications networks in which the total number of machines is much larger than the total number of channels. To minimize the total prediction error of signals collected at the base station in each time slot, we propose the signal-centric predictive polling algorithm. To mitigate the starvation problem, we propose the weighted signal-centric predictive polling algorithm that strikes a good balance between prediction error and fairness. Analytical and simulation results are included to justify the usage of the proposed algorithms.