Delivery route optimization for network efficiency

3 September 2025

Delivery route optimization for network efficiency (DRONE): A two-stage approach balancing efficiency and noise annoyance
A Henry, A Guitart, M Schultz, D Delahaye. Digital Avionics Systems Conference (DASC)

This paper presents a two-stage optimization method improving drone route network efficiency above cities, with a specific focus on balancing delivery performance with noise exposure. The improvement is assessed by three indicators: the mean delivery distance, which reflects travel efficiency and energy use; the number of trajectory intersections, which relates to airspace deconfliction complexity; and the estimated number of highly annoyed individuals due to noise exposure. A twostage approach tackles the trade-off between the three indicators. This approach is based on a Probabilistic Roadmap (PRM) combined with a Simulated Annealing (SA) to improve graph point locations. The combination of the two algorithms iteratively improves the objective function. Our approach is validated through a real-world case study in Paris, France, focusing on medical sample delivery between hospitals and clinics.

Aktuelles

digital twin Cologne-Bonn Airport

Design and integration of new components within the developed virtual airport environment

digital cabin

Design of the experimental setup and conduct of experiments in a digital aircraft cabin

urban air mobility simulator

Enhancement of a simulation environment for urban air traffic under potential emergency conditions

Implications of runway direction changes

This paper presents a runway-aware approach for Singapore Changi Airport that adjusts aircraft trajectories based on dynamic runway capacity, reducing arrival overloads by 81% through modest upstream time shifts without aggressive tactical intervention.

Mitigate air traffic complexity in moving sectors

Moving sectors enable flow-centric management with enhanced airspace capacity, as controller workload is dynamically allocated according to traffic conditions.

Range-angle likelihood maps for indoor positioning

We develop a ResNet-based deep learning approach for indoor aircraft cabin positioning, achieving centimeter-level localization accuracy by transforming range and angle measurements into likelihood grid maps and optimizing model hyperparameters.