Latency-Aware Computation Offloading in Hybrid UAV-Assisted MEC Systems: Time Scheduling and 3D Trajectory Design

The unmanned/uncrewed aerial vehicle (UAV) assisted mobile edge computing (MEC) technology has become a viable and flexible solution for providing computation offloading and energy charging services for ground users, especially in scenarios with terrible direct links. Therefore, latency has become one of the crucial design issues subject to the energy limitations of the UAV and users. Motivated by this, we study a latency-aware air ground hybrid MEC system with an assistant UAV and a ground base station (GBS) to serve and charge multiple users under both the time-division multiple access (TDMA) and non-orthogonal multiple access (NOMA) protocols. The task completion latency minimization problems are formulated by jointly optimizing the time slot scheduling, CPU frequency allocation, UAV's three dimensional (3D) trajectory design, transmit power allocation, as well as the number of required time slots. To address the formulated mixed integer non-convex optimization problems, we introduce an efficient alternating optimization algorithm with a double-loop structure. In the outer loop, we constantly adjust the number of time slots by employing the bisection search method and determine the search range via feasibility check. In the inner loop, we first transform the original subproblem into an equivalent problem that maximizes the minimum computation completion ratio of the users. Then we further deconmpose this transformed problem into four subproblems, which can be solved by a proposed iterative algorithm. Extensive experiments are con ducted to illustrate the efficacy and superiority of the proposed algorithm over the other benchmark schemes in minimizing the task completion latency, particularly in scenarios where the computing resource is limited or the density of users is high.