| Ruifan Liu |
PhD candidate, Cranfield University, School of Aerospace, Transport and Manufacturing, MK43 0AL, Cranfield, UK. | | Hyo-sang Shin |
Professor, Cranfield University, School of Aerospace, Transport and Manufacturing, MK43 0AL, Cranfield, UK. | | Minguk Seo |
Research fellow, Cranfield University, School of Aerospace, Transport and Manufacturing, MK43 0AL, Cranfield, UK. | | Antonios Tsourdos |
Professor, Cranfield University, School of Aerospace, Transport and Manufacturing, MK43 0AL, Cranfield, UK. |
|
| Abstract:
This paper investigates a stochastic route optimization problem for drone delivery systems, with a focus on managing energy risk while optimizing service qualities. In drone delivery, couriers, i.e., small commercial unmanned aerial vehicles (UAVs), face a very limited power and load capacity, which needs to be carefully settled throughout the route optimization process. Additionally, due to its wind-sensitive property, the energy cost of a UAV is fiercely affected by the airflow and might spread across a wide range. To optimize the route while managing its energy use properly, we propose an Energy-aware Planning Framework (EaPF), which is embedded with an energy prediction model and adapts to any existing route optimizer. Instead of sampling using either analytical or numerical formulations, the proposed energy prediction model directly yields the distribution of energy consumption via a Mixture Density Network (MDN). Based on the statistical model, an energy risk criterion and an objective function in the form of expectation are both developed for route optimization. In addition, an event-driven routing simulator is devised with the aim to accommodate the energy model and incorporate it with an optimizer. Finally, we implement the proposed planning framework to the simulation test on a medical delivery mission, demonstrating its superiority in terms of energy risk management and solution qualities.
|
View PDF CEAS-GNC-2022-045