Multi-Access Edge Computing-based Task Offloading for UAV-Assisted Power IoT

Multi-access edge computing (MEC) has emerged as a promising approach to provide emerging delay-sensitive and energy-hungry services and applications for power terminals in power Internet of Things (PIoT) scenarios. However, traditional MEC servers can not be available for power applications in remote areas. Fortunately, unmanned aerial vehicles (UAVs) can be deployed at infrastructure-less environment, such as remote areas, owning to flexible mobility of UAVs. In this paper, a UAV-assisted power IoT system is proposed to execute emerging power services in PIoT scenarios, where the services can be either offloaded to UAV for execution or execution locally. Firstly, we optimize the task offloading and UAV trajectory to minimize the average energy consumption of all power terminals by considering the limited UAV coverage range and maximum tolerable delay of each service. Secondly, considering the coupling between variables and constraints, the primary optimization problem is reformulated as a Markov decision process (MDP), and design a deep reinforcement learning (DRL) algorithm to find the near-optima. Finally, experiment results show that our proposed algorithm can effectively reduce the average energy consumption compared with benchmark solutions.