Accelerating Development in UAV Network Digital Twins with a Flexible Simulation Framework
Md Sharif Hossen, Anil Gurses, Mihail Sichitiu, Ismail Guvenc
TL;DR
The paper addresses the realism and computational bottlenecks in UAV network digital twins by introducing a measurement-calibrated Matlab-based simulation framework that reproduces full-stack UAV wireless-DT performance and is validated against the AERPAW DT. By leveraging 2D Haversine-based geometry, 3D distance calculations, and FSPL with antenna effects, the framework can estimate RSRP and SNR to drive LTE CQI-based throughput predictions. The authors implement an event-driven radio model and mission-planning abstractions (autonomous and fixed trajectories) that align with the DT environment, achieving high similarity in RSRP between simulation and emulation. The framework accelerates development, supports iterative testing before real-world deployment, and is openly available (uavsimframework), with future work including ray tracing, shadowing, interference, and Doppler dynamics to further close the sim-to-real gap.
Abstract
Unmanned aerial vehicles (UAVs) enhance coverage and provide flexible deployment in 5G and next-generation wireless networks. The performance of such wireless networks can be improved by developing new navigation and wireless adaptation approaches in digital twins (DTs). However, challenges such as complex propagation conditions and hardware complexities in real-world scenarios introduce a realism gap with the DTs. Moreover, while using real-time full-stack protocols in DTs enables subsequent deployment and testing in a real-world environment, development in DTs requires high computational complexity and involves a long development time. In this paper, to accelerate the development cycle, we develop a measurement-calibrated Matlab-based simulation framework to replicate performance in a full-stack UAV wireless network DT. In particular, we use the DT from the NSF AERPAW platform and compare its reports with those generated by our developed simulation framework in wireless networks with similar settings. In both environments, we observe comparable results in terms of RSRP measurement, hence motivating iterative use of the developed simulation environment with the DT.