Comparative analysis of UAV simulation platforms: challenges, opportunities, and future directions

Abstract

Unmanned aerial vehicles (UAVs) are deployed across agriculture, disaster response, infrastructure inspection, and military operations, but flight testing is expensive, hazardous, and slow, which has made simulation the default environment for developing and validating UAV algorithms. Dozens of simulation platforms exist, and a substantial secondary literature already compares them; what is missing is a consolidated view across those comparisons. This paper reports a meta-review of nine studies published between 2018 and 2025: six comparative surveys of simulators and agent-based frameworks, two simulator-development papers, and one review of UAV airframes and autonomy components. For each study we reconstruct its scope, the platforms it covers, and the evaluation criteria it applies, and consolidate the results in a single comparison. The surveys converge on a common set of evaluation axes – sensor support, physics fidelity, scalability, usability, and maintenance activity – but weight them differently and share no common benchmark, so their verdicts cannot be compared across publications. Coverage is highly uneven: Gazebo is assessed in five of the nine studies and most other platforms in only one. We argue that the principal gap in UAV simulation is therefore not platform capability but the absence of a shared benchmarking protocol, and set out what such a protocol would have to specify.