Evaluation
Elevator-VIGS
ResearchPublished
Key
innovation
A benchmark dedicated to multi-floor SLAM with inter-floor transitions (elevators), testing floor separation and height estimation.
Category
Evaluation
Abstraction level
Building block
Operation level
Training
How it works
The dataset provides sensor sequences covering movement between floors, enabling evaluation of whether a SLAM system keeps consistent per-floor submaps, detects transitions and correctly estimates height. Editorial note: the dataset's detailed specification (exact sensor suite, authors, license) could not be unambiguously confirmed in available sources as of this write-up — it appears in the context of multi-floor SLAM research (e.g. StrataSLAM).
Problem solved
Existing SLAM benchmarks (e.g. KITTI, TUM RGB-D) do not cover difficult inter-floor transitions or height loss in an elevator, so they do not test the ability to correctly separate floors.