AMB3R-SLAM: Kilometer-Scale Monocular SLAM on One Consumer GPU, 70% Lower ATE
zhenjun_zhao · x · 2026-09-19
The arXiv paper "AMB3R-SLAM: Kilometer-scale SLAM with Hierarchical Backend" presents a real-time monocular SLAM system that reconstructs kilometer-scale trajectories over 10k frames on a single consumer-grade GPU.
- Architecture: a lightweight front-end for low-latency online tracking plus a hierarchical backend enforcing local, mid-level, and global consistency.
- It avoids bundle adjustment's static-world assumption, handling dynamic scenes out of the box, and can take stereo, RGB-D, and LiDAR inputs.
- Results: strong tracking across 9 datasets, cutting absolute trajectory error (ATE) by over 70% vs prior SOTA on VBR and Oxford Spires; with LiDAR, ATE drops to sub-meter on KITTI and VBR.
More from Embodied
- ROS Weekly: ROSCon Global next week in Toronto, InOrbit open-sources Robot Ops — kscottz · 2026-09-19
- MOSS robot demo replays real Jev decisions in simulation, switching tasks from cans to bottles — kamathsblog · 2026-09-19
- Niantic Spatial launches Gaussian Splat Relighting beta for embodied AI simulation — Scobleizer · 2026-09-19
- EliGSiR: continual RGB-D Gaussian Splatting mapping under bounded compute hits 21.52 dB on TUM RGB-D — zhenjun_zhao · 2026-09-19
- SLAMSqueezeBench benchmarks nine SLAM systems under realistic edge compute and memory constraints — zhenjun_zhao · 2026-09-19
- ABC-130K: largest open bimanual teleop dataset with 3,500 hours, headed to CoRL 2026 — pabbeel · 2026-09-19