aiqbench analysis
This research introduces a novel method for 2D point cloud registration in SLAM systems by replacing iterative nearest-neighbor searches with precomputed unsigned distance maps. By leveraging these O(1) lookups and spatial derivatives, the approach significantly reduces computational overhead while maintaining accuracy comparable to standard point-to-plane formulations even under sensor noise. The technique represents a meaningful optimization for real-time laser odometry applications in mobile robotics.
Unsigned Distance Maps on 2D Point Cloud Registration
arXiv:2609.25932v1 Announce Type: new Abstract: 2D point cloud registration arises in laser odometry and Simultaneous Localization and Mapping (SLAM) for mobile robots. Iterative Closest Point (ICP) is one of the most widely used approaches. Still, its iterative procedure recomputes correspondences via nearest-neighbor search at every iteration, whereas correspondence-free alternatives focus on scan-to-map alignment. This paper proposes a 2D point cloud registration approach based on unsigned distance maps, precomputing the Euclidean distance to the nearest reference point, along with its spatial derivatives, over a discrete grid, replacing the per-iteration search with O(1) lookups. Moreover, point-to-point and point-to-plane error formulations are derived on the SE(2) manifold and solved via Gauss-Newton optimization. On a synthetic benchmark and the real-world IILABS 3D dataset, the precomputed point-to-point variant outperforms its analytical counterparts, achieving competitive laser-odometry drift compared to point-to-plane formulations, as the precomputed gradient regularizes correspondences in the presence of sensor noise.
Abstract:2D point cloud registration arises in laser odometry and Simultaneous Localization and Mapping (SLAM) for mobile robots. Iterative Closest Point (ICP) is one of the most widely used approaches. Still, its iterative procedure recomputes correspondences via nearest-neighbor search at every iteration, whereas correspondence-free alternatives focus on scan-to-map alignment. This paper proposes a 2D point cloud registration approach based on unsigned distance maps, precomputing the Euclidean distance to the nearest reference point, along with its spatial derivatives, over a discrete grid, replacing the per-iteration search with O(1) lookups. Moreover, point-to-point and point-to-plane error formulations are derived on the SE(2) manifold and solved via Gauss-Newton optimization. On a synthetic benchmark and the real-world IILABS 3D dataset, the precomputed point-to-point variant outperforms its analytical counterparts, achieving competitive laser-odometry drift compared to point-to-plane formulations, as the precomputed gradient regularizes correspondences in the presence of sensor noise.
8 pages, 0 figures, 4 tables. Accepted to the 9th Iberian Robotics Conference (ROBOT2026), November 18-20, 2026, Barcelona, Spain. Source code: this https URL
Robotics (cs.RO)
arXiv:2609.25932 [cs.RO]
(or arXiv:2609.25932v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2609.25932 arXiv-issued DOI via DataCite (pending registration)
From: Ricardo B. Sousa [view email] [v1] Tue, 22 Sep 2026 09:37:24 UTC (144 KB)
Full text under the publisher's open licence · 198 words
Related coverage
RaDiVe: Robust 4D Radar Odometry with Distance-Bounded NDT and Velocity-Discrepancy Point Uncertainty
arXiv · Jul 31, 2026
Explore From Sketch: Accelerating UAV Exploration in Large-scale Environments with Prior Maps
arXiv · Jun 11, 2026
From Distances to Trajectories: Real-Time Signed Distance Function Mapping and Distance-Accelerated Motion Planning for UAVs
arXiv · Jul 23, 2026
News
Quantum Computing Report · Jul 25, 2026