Drone service response: Spatiotemporal heterogeneity implications

Abstract

Unmanned Aerial Vehicles, often called drones, have rapidly emerged for commercial and personal use in recent years. Drones are a promising and effective transportation mode because they can travel faster than traditional ground-based vehicles, particularly when obstacles limit quick response or in cases of congestion. An important consideration for drones is that travel times are impacted in various ways by real-time local conditions, including weather and terrain. While goods and supplies can be acquired at more traditional outlets (e.g., stores, warehouses, restaurants, hospitals, fire stations, etc.), drones are being increasingly relied upon to extend access, particularly for special services associated with food, drug, and equipment delivery. The reason is that they can reliably access almost anywhere, providing quick response without the need for more expensive (and larger) vehicles that are restricted to congested roadways. How to locate drone base stations and allocate service in order to optimize overall response is a challenging task, especially given spatiotemporal heterogeneity in distributed demand and service response times/costs that can vary over a day. This paper introduces an extension of p-median problem to aid in the deployment of a drone system that accounts for continuous planar travel costs. Results show that drone travel times can be significantly reduced across a region. A key feature in this work is the representation of both demand and flight trajectories across a continuous terrain.