aiqbench analysis
RAIL-5G introduces a rigorous real-world benchmark for AI-driven 5G indoor localization, specifically addressing the severe multipath and non-line-of-sight challenges found in industrial railway environments. The study demonstrates that while self-supervised pretraining significantly reduces labeling costs, robust geometric and fingerprinting methods still struggle to achieve high accuracy under these harsh conditions without extensive supervised training. This dataset provides a critical testbed for evaluating how modern machine learning models generalize from controlled settings to complex, metal-rich industrial deployments.
RAIL-5G: A Challenging Real-World Dataset and Benchmark for AI-Based 5G Indoor Localization
arXiv:2610.25201v1 Announce Type: cross Abstract: Reliable radio-based indoor localization is essential for industrial automation, yet existing methods are often evaluated in environments that do not fully reflect the severe multipath and non-line-of-sight (NLOS) conditions encountered in real deployments. This limits our understanding of how AI, i.e., modern geometric, unsupervised, and semi-supervised localization methods generalize to industrial 5G settings. We present RAIL-5G, a public real-world 5G NR channel impulse response (CIR) benchmark recorded in an operational railway maintenance facility. RAIL-5G covers an 811m$^2$ indoor area with eight distributed radio units, robot-mounted user equipment (UE) trajectories, high-accuracy iGPS reference positions, and severe metal-induced multipath caused by train compartments and maintenance infrastructure. We benchmark geometric time-difference of arrival (TDoA), supervised fingerprinting, unsupervised channel charting, and self-supervised wireless pretraining methods. Our results highlight the difficulty of RAIL-5G: on the full dataset, TDoA reaches 7.02m mean average error (MAE), the best unsupervised method reaches 2.46m, and supervised fingerprinting achieves 0.52m. Self-supervised pretraining reduces the labeling effort by up to one order of magnitude, but the remaining gap to supervised learning shows that robust AI-based localization under severe multipath remains an open challenge. The dataset and code are available at https://iis.fraunhofer.de/rail-5g.
Abstract:Reliable radio-based indoor localization is essential for industrial automation, yet existing methods are often evaluated in environments that do not fully reflect the severe multipath and non-line-of-sight (NLOS) conditions encountered in real deployments. This limits our understanding of how AI, i.e., modern geometric, unsupervised, and semi-supervised localization methods generalize to industrial 5G settings. We present RAIL-5G, a public real-world 5G NR channel impulse response (CIR) benchmark recorded in an operational railway maintenance facility. RAIL-5G covers an 811m$^2$ indoor area with eight distributed radio units, robot-mounted user equipment (UE) trajectories, high-accuracy iGPS reference positions, and severe metal-induced multipath caused by train compartments and maintenance infrastructure. We benchmark geometric time-difference of arrival (TDoA), supervised fingerprinting, unsupervised channel charting, and self-supervised wireless pretraining methods. Our results highlight the difficulty of RAIL-5G: on the full dataset, TDoA reaches 7.02m mean average error (MAE), the best unsupervised method reaches 2.46m, and supervised fingerprinting achieves 0.52m. Self-supervised pretraining reduces the labeling effort by up to one order of magnitude, but the remaining gap to supervised learning shows that robust AI-based localization under severe multipath remains an open challenge. The dataset and code are available at this https URL.
Signal Processing (eess.SP); Machine Learning (cs.LG)
arXiv:2610.25201 [eess.SP]
(or arXiv:2610.25201v1 [eess.SP] for this version)
https://doi.org/10.48550/arXiv.2610.25201 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 287 words
Related coverage
Multi-viewpoint Geo-localization with Event Cameras
arXiv · Sep 21, 2026
Fault Lines: Navigating Ethics and Responsible AI Where National Policy Meets Local Practice in Public Sector Transformation
arXiv · Jun 12, 2026
X$^2$Localizer: Cross-grained Alignment for Progressive Cross-view Video Geo-localization
arXiv · Aug 18, 2026
Hilti-Trimble-Oxford Dataset: 360 Visual-Inertial Benchmark with Floor Plan Priors for SLAM and Localization
arXiv · Oct 22, 2026