One of the most promising use cases of 5G/IMT2020 is the unmanned aerial vehicle (UAV).
Due to their small size, the UAVs are resource constraint devices. To this end, this paper proposes an
offloading algorithm for UAVs to assist in the execution of computationally intensive tasks. The proposed
algorithm provides two UAV offloading methods. The first offloading method is the air-offloading, where a
UAV can offload its computing tasks to nearby UAVs that have available computing and energy resources.
The second offloading method is the ground-offloading, which enables the offloading of tasks to an edge
cloud server from the multi-level edge cloud units connected to ground stations. The proposed algorithm is
energy- and latency-aware, i.e., it selects the execution device and the offloading method based on the latency
and energy constraints. The intensive algorithm simulation over reliable conditions for various scenarios with
different cases for each scenario is conducted and results are presented.