Joint Laser Charging and DBS Placement for Drone-Assisted Edge Computing

Abstract

Deploying a drone-mounted base-station (DBS) to assist mobile edge computing can empower 5&#x00A0;G and beyond networks with additional flexibility and maneuverability, and laser charging can potentially extend the DBS&#x0027;s service time. A laser charging DBS framework is proposed in which a DBS is provisioning services for user equipments (<i>UEs</i>) on the ground and harvesting energy transmitted from the laser charging station mounted on the macro base station (<i>MBS</i>). Both MBS and DBS are equipped with servers. UEs can offload their tasks to either the MBS or the DBS. The DBS is to be placed at the optimal location to provide uplink communications and computing services for the ground UEs in each time slot. We thus formulate the joint user <u>A</u>sso<u>C</u>iation bandwidth and <u>C</u>omputation <u>R</u>eso<u>U</u>rce <u>A</u>ssignment <u>L</u>aser charing (<i>ACCRUAL</i>) problem to jointly maximize the DBS service time and minimize the task completion time for all UEs. Since ACCRUAL is a mixed integer nonlinear problem, we decompose it into two sub-problems: the joint UE association computing <u>R</u><u>E</u>source <u>A</u>nd bandwidth allocation <u>P</u>roblem (<i>REAP</i>) and the DBS placement problem. An iterative algorithm is employed to solve the REAP problem and a placement algorithm based on counting sort is used to tackle the DBS placement problem. The performance of our algorithm is superior to the greedy algorithm and equally shared resource allocation algorithm upon which the total UE task offloading completion time is improved by more than 9&#x0025; as compared to the greedy algorithm and the total DBS service time is improved by 20&#x0025; when the laser power is 200&#x00A0;w and 40&#x0025; of the DBS service time is improved when the laser power is 400&#x00A0;w.