Joint Computation Offloading and Trajectory Design for Aerial Computing

Abstract

Aerial computing is a key enabling technique to provide seamless telecommunication and IT services for emerging computation-craving applications. As a drone employed in an aerial computing platform is limited in computation, caching, communication, and control capabilities, researchers have utilized multiple drones for efficient multi-modal multi-task processes, which also bring great challenges in massive data transmission and dynamic network management. Motivated by recent advancement of device-to-device communication and software defined networking technologies, we propose an air-ground integrated aerial computing framework where a cloudlet server and multiple mobile edge servers are mounted on drones to provide reliable and efficient edge computing services for ground devices. The joint implementation of computation offloading and trajectory design for the proposed framework is investigated. Extensive numerical results demonstrate that our aerial computing framework achieves better performance in terms of average task completion time than traditional ones.