Reproduce “LIMOncello: Iterated Error-State Kalman Filter on the SGal(3) Manifold for Fast LiDAR-Inertial Odometry” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Carlos Pérez-Ruiz, Joan Solà

LIMOncello introduces a LiDAR-Inertial Odometry system utilizing an Iterated Error-State Kalman Filter on the SGal(3) manifold and an efficient i-Octree mapping backend. The system maintained stable trajectories in geometrically degenerate environments, such as a 400m tunnel where other methods failed, and showed an average mapping runtime of 14.6 ms.

Key insights

  • Employing the SGal(3) manifold for state representation intrinsically couples spatial and temporal quantities, leading to more consistent motion propagation, particularly under low LiDAR observability.
  • Simplifying the i-Octree by omitting continuous rebalancing and reallocation steps results in a lightweight, incremental mapping backend that maintains efficiency and stable memory usage for real-time operation.
  • A carefully designed discrete-time SGal(3)-based formulation can achieve competitive accuracy in feature-rich environments and improved robustness in challenging degenerate conditions.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.