Fine-Grained Task Offloading for UAV via MEC-Enabled Networks

The ground mobile edge computing (MEC) system has been effectively utilized to undertake the computation-intensive tasks offloaded from unmanned aerial vehicles (UAVs), which has significantly mitigated the aerial calculation pressure. However, some external factors, like environmental conditions and the distribution of MEC servers can deeply affect the performance of offloading algorithms. In this paper, an enhanced offloading algorithm is proposed to minimize the the completion time. For the sake of practice, the air-to-ground (A2G) channel model is rebuilt with the line of sight (LoS)/non-line of sight (NLoS) status considered. Furthermore, the boundary of effective offloading area with dense MEC servers is denoted by the round margin raised. Within the round margin, UAV offloads its calculation to the ground and plans its trajectory simultaneously. Outside the round margin, UAV flies along the straight path with maximum speed, which avoids inefficient operations within the sparse and deviated area. Simulation results show that the proposed scheme is validated with better performance. Moreover, the differences of offloading effectiveness under different conditions of sparsity or deviation provide potential instructions for future trade-off research between offloading and local computing.