PNE-SGAN: Probabilistic NDT-Enhanced Semantic Graph Attention Network for LiDAR Loop Closure Detection

2025-04-11

AI Summary

LiDAR-based loop closure detection (LCD) in SLAM suffers from insufficient robustness and accuracy under dynamic objects, sensor noise, and viewpoint variations. To address this, we propose a spatiotemporal graph neural network framework that jointly exploits geometric and semantic cues. Specifically, we introduce the NDT covariance matrix as a discriminative geometric node feature—a novel design enabling robust geometric representation. We further construct a semantic graph and model its structural dependencies using a Graph Attention Network (GAT). Additionally, we develop a probabilistic temporal similarity model grounded in Hidden Markov Models (HMMs) and Bayesian filtering, enhanced with forward–backward smoothing to mitigate loop ambiguity. Evaluated on KITTI sequences 00 and 08, our method achieves mean average precision of 96.2% and 95.1%, respectively—outperforming state-of-the-art approaches, particularly in challenging bidirectional loop scenarios.

Technology Category

• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Intelligent Robots: State Estimation
• Machine Learning: Graph-based Machine Learning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

LiDAR loop closure detection (LCD) is crucial for consistent Simultaneous Localization and Mapping (SLAM) but faces challenges in robustness and accuracy. Existing methods, including semantic graph approaches, often suffer from coarse geometric representations and lack temporal robustness against noise, dynamics, and viewpoint changes. We introduce PNE-SGAN, a Probabilistic NDT-Enhanced Semantic Graph Attention Network, to overcome these limitations. PNE-SGAN enhances semantic graphs by using Normal Distributions Transform (NDT) covariance matrices as rich, discriminative geometric node features, processed via a Graph Attention Network (GAT). Crucially, it integrates graph similarity scores into a probabilistic temporal filtering framework (modeled as an HMM/Bayes filter), incorporating uncertain odometry for motion modeling and utilizing forward-backward smoothing to effectively handle ambiguities. Evaluations on challenging KITTI sequences (00 and 08) demonstrate state-of-the-art performance, achieving Average Precision of 96.2% and 95.1%, respectively. PNE-SGAN significantly outperforms existing methods, particularly in difficult bidirectional loop scenarios where others falter. By synergizing detailed NDT geometry with principled probabilistic temporal reasoning, PNE-SGAN offers a highly accurate and robust solution for LiDAR LCD, enhancing SLAM reliability in complex, large-scale environments.

Problem

• Enhances LiDAR loop closure detection robustness and accuracy
• Overcomes coarse geometry and noise sensitivity in existing methods
• Integrates probabilistic temporal filtering for handling ambiguities

Innovation

• NDT covariance matrices enhance semantic graphs
• Graph Attention Network processes geometric features
• Probabilistic temporal filtering handles ambiguities