Robustness of LiDAR-Based Pose Estimation: Evaluating and Improving Odometry and Localization Under Common Point Cloud Corruptions

2024-09-17

AI Summary

This work addresses the insufficient robustness of LiDAR pose estimation systems under point cloud degradation induced by adverse weather conditions (e.g., rain, fog). To systematically evaluate performance degradation, we introduce the first standardized point cloud corruption benchmark tailored for LiDAR odometry and localization tasks, assessing five state-of-the-art methods under 18 realistic corruption types. We find that odometry is highly sensitive to local structural distortions—inducing over 80% relative pose error increase—whereas localization exhibits greater inherent robustness. Building on these insights, we propose a dual-path robustness enhancement framework: (1) a point cloud denoising path leveraging PVD/PCN reduces errors from noise-type corruptions by 37% on average; (2) an adversarial retraining path using synthetically corrupted data improves model robustness across 12 corruption types by 52% on average.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Accurate and reliable pose estimation, i.e., determining the precise position and orientation of autonomous robots and vehicles, is critical for tasks like navigation and mapping. LiDAR is a widely used sensor for pose estimation, with odometry and localization being two primary tasks. LiDAR odometry estimates the relative motion between consecutive scans, while LiDAR localization aligns real-time scans with a pre-recorded map to obtain a global pose. Although they have different objectives and application scenarios, both rely on point cloud registration as the underlying technique and face shared challenges of data corruption caused by adverse conditions (e.g., rain). While state-of-the-art (SOTA) pose estimation systems achieved high accuracy on clean data, their robustness to corrupted data remains unclear. In this work, we propose a framework to systematically evaluate five SOTA LiDAR pose estimation systems across 18 synthetic real-world point cloud corruptions. Our experiments reveal that odometry systems degrade significantly under specific corruptions, with relative position errors increasing from 0.5% to more than 80%, while localization systems remain highly robust. We further demonstrate that denoising techniques can effectively mitigate the adverse effects of noise-induced corruptions, and re-training learning-based systems with corrupted data significantly enhances the robustness against various corruption types.

Problem

• Evaluating robustness of LiDAR pose estimation under data corruption.
• Improving odometry and localization accuracy in adverse conditions.
• Assessing impact of synthetic real-world corruptions on SOTA systems.

Innovation

• Systematic evaluation of LiDAR pose estimation robustness
• Denoising techniques mitigate noise-induced corruptions effectively
• Re-training with corrupted data enhances system robustness