RootSLAM: square-root bundle adjustment makes visual-inertial SLAM up to 20x faster

zhenjun_zhao · x · 2026-10-17

• Positioning: Efficient SLAM for small, low-power robots, MR, autonomous driving, and drones — learned SLAM is accurate but compute-heavy, classical geometry stays efficient.
• Method: RootSLAM builds a full multi-camera visual-inertial SLAM system on Basalt, adapting square-root bundle adjustment from structure-from-motion via a novel reformulation that lets the global optimizer run in single precision, doubling performance; adds map database, loop closing, and global BA modules.
• Results: State-of-the-art accuracy while running up to 20x faster; across 95 sequences mean ATE drops from Basalt's 95cm to 12.4cm, versus 94.3cm for OKVIS2.
• Open source: full implementation including map management, loop detection, pose-graph optimizer, and global mapping will be released.

Original post →

More from Embodied

• Contact-UAN: MIT's residual world model fixes humanoid sim-to-real with 6.5 minutes of data — kevin_zakka · 2026-10-17
• ROS 2 LiDAR localization runs 390 m outdoors on handheld Livox MID-360 — rsasaki0109 · 2026-10-17
• Embodied Astra tops RoboMME at 92.1% — beating humans by just keeping full visual history in context — hankyang94 · 2026-10-17
• ROS weekly: ROSCon China kicks off in Shanghai, ROS By-The-Bay set for Oct 22 at Intrinsic — kscottz · 2026-10-17
• Simulated robotics eval: Claude Opus 5.5 preferred over Opus 5 in 14 of 16 comparisons — qinzytech · 2026-10-17
• FoundationVIO fuses IMU with foundation-model predictions to fix metric scale in feed-forward SLAM — zhenjun_zhao · 2026-10-17

Embodied channel →