IPNL — Intelligent Positioning and Navigation LaboratoryUrbanLoco Dataset
dataset CC BY-NC-SA 4.0UrbanLoco is a multi-sensor dataset for benchmarking mapping and localization algorithms in highly urbanized environments. It contains 13 trajectories totaling over 40 km across San Francisco and Hong Kong — urban canyons, bridges, tunnels, and sharp turns — with synchronized LiDAR, cameras (360° in SF; fish-eye sky-camera in HK), IMU, and GNSS measurements. Ground truth is provided by a NovAtel SPAN CPT 7 system with RTK corrections. The dataset accompanies the ICRA 2020 paper by Wen, Zhou, Zhang, Fahandezh-Saadi, Bai, Zhan, Tomizuka, and Hsu.
Repository: https://github.com/weisongwen/UrbanLoco
Citation: Wen, W., et al. (2020). UrbanLoco: A full sensor suite dataset for mapping and localization in urban scenes. ICRA.