A Mean-Field Framework for Energy-Efficient Power Control in Massive IoT Environments

This paper presents an uplink energy-efficient power control mechanism for massive cellular Internet of Things (IoT) devices using a Mean-Field Game (MFG) approach. In this setting, IoT devices are clustered around closed access small base stations supporting massive connectivity, in order to avoid data traffic congestion. Each IoT device adapts its transmit power to its energy level and selfishly attempts to satisfy its quality of service expressed in terms of signal to interference plus noise ratio, while reducing the power consumption. The power control is first modeled as a differential game then extend to a MFG considering two cases: Large scale network and ultra dense IoT network. The mean field interference are derived using stochastic geometry analysis. Therefore, the IoT devices can predict their optimal transmit power policies based only on their initial energy distribution. A finite difference algorithm is then developed to obtain the mean field equilibrium. The simulations illustrate the optimal transmit power and the mean field at the equilibrium for both orthogonal and non orthogonal multiple access.