# A New Approach to 3D ICP Covariance Estimation

arXiv: https://arxiv.org/abs/1909.05722 · HTML: https://mlsanity.com/paper/1909.05722/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Sep 2019.

## Built on by (most cited first, 9 of 9)

- [Learning-based Localizability Estimation for Robust LiDAR Localization](https://mlsanity.com/paper/2203.05698/) (2022)
- [Efficient LiDAR Odometry for Autonomous Driving](https://mlsanity.com/paper/2104.10879/) (2021)
- [Stein ICP for Uncertainty Estimation in Point Cloud Matching](https://mlsanity.com/paper/2106.03287/) (2021)
- [Deep Bayesian ICP Covariance Estimation](https://mlsanity.com/paper/2202.11607/) (2022)
- [Learning minimal volume uncertainty ellipsoids](https://mlsanity.com/paper/2405.02441/) (2024)
- [ICET Online Accuracy Characterization for Geometry-Based Laser Scan Matching](https://mlsanity.com/paper/2306.08690/) (2023)
- [Object Pose Distribution Estimation for Determining Revolution and Reflection Uncertainty in Point Clouds](https://mlsanity.com/paper/2512.07211/) (2025)
- [Benchmarking Classical and Learning-Based Multibeam Point Cloud Registration](https://mlsanity.com/paper/2405.06279/) (2024)
- [Neural Error Covariance Estimation for Precise LiDAR Localization](https://mlsanity.com/paper/2501.02558/) (2025)

## Abstract (arXiv)

In mobile robotics, scan matching of point clouds using Iterative Closest Point (ICP) allows estimating sensor displacements. It may prove important to assess the associated uncertainty about the obtained rigid transformation, especially for sensor fusion purposes. In this paper we propose a novel approach to 3D uncertainty of ICP that accounts for all the sources of error as listed in Censi's pioneering work [1], namely wrong convergence, underconstrained situations, and sensor noise. Our approach builds on two facts. First, the uncertainty about the ICP's output fully depends on the initialization accuracy. Thus speaking of the covariance of ICP makes sense only in relation to the initialization uncertainty, which generally stems from odometry errors. We capture this using the unscented transform, which also reflects correlations between initial and final uncertainties. Then, assuming white sensor noise leads to overoptimism as ICP is biased owing to e.g. calibration biases, which we account for. Our solution is tested on publicly available real data ranging from structured to unstructured environments, where our algorithm predicts consistent results with actual uncertainty, and compares favorably to previous methods.