How Realistic Are UAV Routing Studies? An Assessment of Operational Realism, Validation Maturity, and the Path to Deployable Missions
This systematic review of 215 UAV routing studies reveals a critical gap between algorithmic sophistication and deployment readiness, highlighting that the majority of research lacks operational realism, physical validation, and reproducibility despite advances in modeling complexity.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine a small, silent aircraft hovering over a city, tasked with delivering a package, inspecting a power line, or mapping a forest fire. For this machine to succeed, it must do more than just fly from point A to point B; it must navigate a world that is constantly changing. The wind might shift, the battery might drain faster than expected due to a heavy load, or a new obstacle might appear. To plan these flights, researchers build complex computer models that calculate the best possible path. These models are the digital blueprints for real-world missions, designed to tell the drone exactly where to go, how fast to fly, and when to land. The hope is that a perfect plan on a screen will translate into a perfect flight in the sky. However, there is a critical question that often goes unasked: just how much of the real world do these digital blueprints actually capture?
A recent systematic review by Mehmet Çevik and Latif Tibet Aktaş, published in September 2026, set out to answer this question by examining the last decade of research into drone routing. The authors gathered and analyzed 215 scientific studies to see how closely the computer simulations matched the messy, unpredictable reality of flying a physical machine. They looked at whether researchers were accounting for the specific physics of the drone, such as how a heavy battery or strong wind changes energy use, and whether they were testing their ideas with actual flying machines or just running numbers on a computer. The review also checked how well these studies were validated, meaning whether the proposed routes were ever tried out in the real world or if they remained purely theoretical exercises.
The researchers found a striking gap between the sophistication of the computer algorithms and the reality of the missions they were meant to guide. While the mathematical methods used to plan these routes have become incredibly advanced, the models often ignore the physical constraints that would cause a real drone to fail. In the 215 studies reviewed, more than two-thirds relied on simplified assumptions about energy, treating the drone's battery like a simple timer or a fixed distance limit rather than a complex system that drains differently depending on speed, weight, and weather. Furthermore, the vast majority of these studies, about 71 percent, treated the environment as static and predictable, ignoring the fact that real missions often involve sudden changes like moving targets, shifting weather, or unexpected mechanical failures.
Perhaps the most revealing finding concerned how these ideas were tested. The review showed that the field is heavily concentrated on computer simulations. About 67 percent of the studies evaluated their methods using synthetic data or standard benchmarks created entirely within a computer. Only a tiny fraction, roughly 10 percent, involved flying an actual drone to test the plan, and a mere 2 percent reached the stage of a full operational deployment in a real-world setting. This means that for every hundred studies published, only two had been proven to work outside of a simulation. The authors noted that even when researchers used real-world data, such as actual city maps or real weather records, they often still ran the entire test inside a computer, which does not guarantee that the drone could physically execute the route.
The study also highlighted that the complexity of the algorithm did not necessarily lead to better real-world readiness. Whether researchers used traditional mathematical methods or newer artificial intelligence techniques, the level of physical validation remained low. In fact, the most advanced routing architectures, which tried to coordinate multiple drones or handle complex, dynamic tasks, were almost exclusively tested in simulation. The review identified a specific disconnect: researchers were becoming better at solving complex puzzles on paper but were not keeping pace with the need to prove those solutions worked on actual hardware. For instance, while many studies claimed to address the challenges of charging drones or coordinating fleets, very few of these complex scenarios were ever demonstrated with physical machines.
To bridge this gap, the authors suggest a shift in how the field approaches its work. They argue that future research must prioritize "physics-aware" modeling, where the computer understands the drone as a physical object subject to gravity, wind, and battery chemistry, rather than just a point moving on a map. They also call for "progressive validation," a step-by-step process where a method is tested first in simulation, then with real data, then in a controlled flight test, and finally in a real mission. The review emphasizes that the goal should not just be to find the shortest path or the fastest algorithm, but to create routing methods that are robust enough to be flown by a real drone in a changing world. Until researchers move beyond the comfort of the computer screen and embrace the unpredictability of the sky, the promise of autonomous drone missions will remain largely unfulfilled.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.