Reproducibility remains a major challenge in vehicular network research, with many studies relying on proprietary simulation environments that are difficult to compare or reproduce.

To support open and reproducible research, the 6G-TWIN project has released NeversNet5G, a new city-scale dataset, now publicly available on Zenodo.

Generated through a SUMO–Simu5G co-simulation over the real road network of Nevers (France), the dataset includes 390 vehicle trajectories, 19 real-world base station locations, four network load scenarios, and more than 242 million synchronized records combining vehicle mobility (position, speed, road segment) with network measurements such as SINR, CQI, throughput, and latency.

This dataset was developed by our partners at the Université Bourgogne Europe, and we’re pleased to make it openly available to the research community.

NeversNet5G is designed to support research on Network Digital Twins, network-aware routing, QoS prediction, handover management, and AI-driven optimisation for connected mobility.

Dataset: https://doi.org/10.5281/zenodo.21225461

A companion scientific paper describing the dataset and its methodology is currently in preparation and will be submitted for publication soon.

We encourage the research community to use the dataset and share feedback to help improve future releases.