Optimal Trajectory Planning and Task Assignment for UAV-assisted Fog Computing

Fog computing is an emerging distributed computing model for the realization of the Internet of Things (IoT). It extends computing and caching functions to the edge of wireless networks. Unmanned Aerial Vehicles (UAVs) provide effective support for fog computing, as UAVs can not only act as a relay between mobile users and physically remote edge devices to avoid costly long-range wireless communications, but also are equipped with computing facilities that can take over certain tasks. In this paper, we aim to optimize the energy efficiency of a fog computing system assisted by a single UAV, by planning the trajectories of the UAV and assigning computing tasks to different devices including the UAV itself. We propose two algorithms based on the classical Ant Colony and Particle Swarm Optimization techniques, and solve the problem by continuous convex approximation. Different from most existing studies where the trajectories are assumed to be straight lines, we account for the effect of obstacles, such as buildings, and deliberately avoid them during the trajectory planning phase. We demonstrate, through extensive simulation experiments, that our proposed approach can achieve significantly better energy efficiency compared to existing benchmark algorithms.