부산대학교 도서관

Due to flexible deployment and low cost, an unmanned aerial relaying (UAR) system has been considered as a promising aerial communication platform to complement current terrestrial infrastructure in some scenarios like temporary hotspots, disaster areas, and complex terrains. However, practical UAR systems may not be able to provide ubiquitous and high-quality wireless connections, mainly due to the limited on-board energy of unmanned aerial vehicles (UAVs). Consequently, the UAVs are unable to hover in the sky for a long time, thus having to land frequently for energy replenishment. To save UAVs' energy, we propose a joint communication and placement design for energy-efficient communications in a multi-hop UAR system. Specifically, we formulate the Energy Consumption Rate minimization problem, called ECR, by jointly considering routing, data rate, transmission power, and UAV position. Since ECR is a non-convex nonlinear programming, we propose an iterative optimization method to solve ECR iteratively by first reducing ECR into two sub-problems ECR-C and ECR-P and then solving them sequentially at an iteration. Specifically, we develop an $\epsilon$-bounded approximation algorithm to solve ECR-C with a performance guarantee, and reformulate ECR-P as a global consensus optimization problem solved by an ADMM (alternating direction method of multipliers) method. Particularly, all UAVs locally update their own positions in parallel, reducing the computation and implementation complexity significantly. Extensive simulation results show the energy efficiency improvement by up to 20% compared with other existing algorithms.