Two-Way-Ranging-based Cooperative Localization with an Open UWB Dataset from Bern-University (TWR-CLOUD-Bern)

TWR-CLOUD-Bern is an open ultra-wideband (UWB) two-way ranging (TWR) dataset for cooperative localization. It was collected with 40 Qorvo DWM3001C nodes over 12 scans covering more than 500 m² at the University of Bern, and provides ground-truth coordinates, node-to-node time-of-flight/range measurements, reception statistics, and protocol-stage statistics.

The release includes four Python/Pandas measurement and statistics tables, JSON benchmark scenarios, Python and Java parsers, a deterministic release-generation script, and the official score_twr_cloud_bern.py scorer for comparable localization-output evaluation. Eleven scans support training and ranging-model development, while one independent scan supports held-out evaluation under anchor-free cooperative, anchor-aided cooperative, and anchor-only non-cooperative positioning modes.

The complete technical documentation is available at twr-cloud.inf.unibe.ch.