Delivery with UAVs: a simulated dataset via ATS

We consider a delivery food service operated by Unmanned Aerial Vehicles (UAVs). Due to the absence of a dataset on UAVs deliveries in the literature, and since it is not possible to perform real tests, we create a dataset using an open Air Traffic Simulator (ATS). Precisely, we converted a set of food deliveries operated by wheeled vehicles, proposed in the literature [1], into a set of simulated UAVs deliveries. For each delivery, we ran a UAV flight from the source to the destination. The results showed that, as expected, the UAV's course is shorter than the vehicle trajectory on the ground because the UAV follows an Euclidean path. Following that path, UAVs can be 5 to 8 times faster than wheeled vehicle, in absence of wind. Highly important, the ATS simulator allows to take care of the wind impact in a realistic way. Tailwind increases UAVs speed which becomes up to 10 times faster than the wheeled vehicles, whereas the headwind and crosswind slowdown the UAVs as the traffic slowdown the wheeled vehicles. Our work proves that air traffic simulators pave the way for realistic simulations of UAVs systems.