# Learning-based Localizability Estimation for Robust LiDAR Localization

arXiv: https://arxiv.org/abs/2203.05698 · HTML: https://mlsanity.com/paper/2203.05698/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 36 references that are in the map.
- First posted on arXiv 11 Mar 2022.

## Builds on (most cited first, 14 of 14)

- [PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation](https://mlsanity.com/paper/1612.00593/) (2016)
- [4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks](https://mlsanity.com/paper/1904.08755/) (2019)
- [Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloud](https://mlsanity.com/paper/2003.01251/) (2020)
- [Road Extraction by Deep Residual U-Net](https://mlsanity.com/paper/1711.10684/) (2017)
- [LO-Net: Deep Real-time Lidar Odometry](https://mlsanity.com/paper/1904.08242/) (2019)
- [CERBERUS: Autonomous Legged and Aerial Robotic Exploration in the Tunnel and Urban Circuits of the DARPA Subterranean Challenge](https://mlsanity.com/paper/2201.07067/) (2022)
- [Self-supervised Learning of LiDAR Odometry for Robotic Applications](https://mlsanity.com/paper/2011.05418/) (2020)
- [CELLO-3D: Estimating the Covariance of ICP in the Real World](https://mlsanity.com/paper/1810.01470/) (2018)
- [Graph-based Multi-sensor Fusion for Consistent Localization of Autonomous Construction Robots](https://mlsanity.com/paper/2203.01389/) (2022)
- [A New Approach to 3D ICP Covariance Estimation](https://mlsanity.com/paper/1909.05722/) (2019)
- [CorAl -- Are the point clouds Correctly Aligned?](https://mlsanity.com/paper/2109.09820/) (2021)
- [On the Covariance of ICP-based Scan-matching Techniques](https://mlsanity.com/paper/1410.7632/) (2014)
- [Estimating Motion Uncertainty with Bayesian ICP](https://mlsanity.com/paper/2004.07973/) (2020)
- [PointNetKL: Deep Inference for GICP Covariance Estimation in Bathymetric SLAM](https://mlsanity.com/paper/2003.10931/) (2020)

## Built on by (most cited first, 11 of 11)

- [G3Reg: Pyramid Graph-based Global Registration using Gaussian Ellipsoid Model](https://mlsanity.com/paper/2308.11573/) (2023)
- [DOA: A Degeneracy Optimization Agent with Adaptive Pose Compensation Capability based on Deep Reinforcement Learning](https://mlsanity.com/paper/2507.19742/) (2025)
- [Dense 3D Reconstruction Through Lidar: A Comparative Study on Ex-vivo Porcine Tissue](https://mlsanity.com/paper/2401.10709/) (2024)
- [Holistic Fusion: Task- and Setup-Agnostic Robot Localization and State Estimation with Factor Graphs](https://mlsanity.com/paper/2504.06479/) (2025)
- [Implicit neural representation for change detection](https://mlsanity.com/paper/2307.15428/) (2023)
- [Prepared for the Worst: A Learning-Based Adversarial Attack for Resilience Analysis of the ICP Algorithm](https://mlsanity.com/paper/2403.05666/) (2024)
- [Neural Error Covariance Estimation for Precise LiDAR Localization](https://mlsanity.com/paper/2501.02558/) (2025)
- [CompSLAM: Complementary Hierarchical Multi-Modal Localization and Mapping for Robot Autonomy in Underground Environments](https://mlsanity.com/paper/2505.06483/) (2025)
- [DCReg: Decoupled Characterization for Efficient Degenerate LiDAR Registration](https://mlsanity.com/paper/2509.06285/) (2025)
- [Towards Robust LiDAR Localization: Deep Learning-based Uncertainty Estimation](https://mlsanity.com/paper/2509.18954/) (2025)
- [LF-GICP: Parameter-Free Degeneracy-Aware LiDAR Odometry via a Voxel-Normal Localizability Field](https://mlsanity.com/paper/2608.19522/) (2026)

## Abstract (arXiv)

LiDAR-based localization and mapping is one of the core components in many modern robotic systems due to the direct integration of range and geometry, allowing for precise motion estimation and generation of high quality maps in real-time. Yet, as a consequence of insufficient environmental constraints present in the scene, this dependence on geometry can result in localization failure, happening in self-symmetric surroundings such as tunnels. This work addresses precisely this issue by proposing a neural network-based estimation approach for detecting (non-)localizability during robot operation. Special attention is given to the localizability of scan-to-scan registration, as it is a crucial component in many LiDAR odometry estimation pipelines. In contrast to previous, mostly traditional detection approaches, the proposed method enables early detection of failure by estimating the localizability on raw sensor measurements without evaluating the underlying registration optimization. Moreover, previous approaches remain limited in their ability to generalize across environments and sensor types, as heuristic-tuning of degeneracy detection thresholds is required. The proposed approach avoids this problem by learning from a collection of different environments, allowing the network to function over various scenarios. Furthermore, the network is trained exclusively on simulated data, avoiding arduous data collection in challenging and degenerate, often hard-to-access, environments. The presented method is tested during field experiments conducted across challenging environments and on two different sensor types without any modifications. The observed detection performance is on par with state-of-the-art methods after environment-specific threshold tuning.