Answers are grounded in this paper and cite their sources.
The Spot Autowalk Long-Term (SALT) Dataset was collected by performing repeated Autowalks with a Boston Dynamics Spot robot across three different scenarios, including both indoor and outdoor environments. The main contribution is 217 traversals of an outdoor Autowalk around the Ontario Tech University campus over a 16 month period. The recorded data includes point clouds, local elevation and occupancy grids, dynamic object and staircase detection, odometry data, joint states, and battery conditions. In total, the dataset contains 286 traversals and 193 km of walking over a period of nearly 2 years. The data is presented in multiple formats for ease of use, including: native format, a human-readable format, a KITTI Odometry format including pseudo ground truth poses, and an MCAP format for use with ROS2. This dataset aims to enable researchers to develop new algorithms for long-term mapping applications. The SALT dataset can be found here https://doi.org/10.5281/zenodo.20670050
No takes yet. Share an insight, caveat, or question.
Baird et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: