Energy-Efficient Associations for IoT networks with UAV: A Regret Matching Based Approach

Abstract

Energy-Efficiency (EE) is an important issue in the IoT system. Unmanned Aerial Vehicles (UAVs) have been used as aggregators to collect data from IoT ground devices, and provide an energy-efficient and cost-effective solution. In this paper, we aim at maximizing the overall EE of the IoT network, by finding to most suitable IoT-UAV association. We formulate the problem as a non-cooperative game where IoT players choose the UAVs that minimize their transmit power by learning their best strategy using an approach based on regret-matching learning. Simulations results show a fast convergence to an optimal solution that provides a low average total transmit power and maximizes the IoT system's overall EE.