KN-LIO: Geometric Kinematics and Neural Field Coupled LiDAR-Inertial Odometry

2025-01-08

AI Summary

To address the insufficient coupling between geometric estimation and neural representation in LiDAR-inertial odometry (LIO) for simultaneous localization and dense mapping under high-dynamic conditions, this paper proposes KN-LIO—a tightly coupled framework integrating kinematic geometry and neural implicit fields. Methodologically, it introduces a novel dual-paradigm architecture combining semi-coupled and tightly coupled designs, supporting asynchronous multi-LiDAR inputs. It unifies iterative error-state Kalman filtering, online signed distance function (SDF) neural field decoding, and joint geometric-neural state propagation and measurement fusion. Experiments on multiple high-dynamic datasets demonstrate state-of-the-art (SOTA) or superior pose estimation accuracy; dense mapping achieves significantly higher fidelity than pure LiDAR-based methods, with minimal information loss. KN-LIO thus enables both high-precision localization and generalizable neural radiance field (NeRF)-enabled mapping.

Technology Category

• Intelligent Robots: State Estimation
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Search and Optimization: Learning to Search

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: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Recent advancements in LiDAR-Inertial Odometry (LIO) have boosted a large amount of applications. However, traditional LIO systems tend to focus more on localization rather than mapping, with maps consisting mostly of sparse geometric elements, which is not ideal for downstream tasks. Recent emerging neural field technology has great potential in dense mapping, but pure LiDAR mapping is difficult to work on high-dynamic vehicles. To mitigate this challenge, we present a new solution that tightly couples geometric kinematics with neural fields to enhance simultaneous state estimation and dense mapping capabilities. We propose both semi-coupled and tightly coupled Kinematic-Neural LIO (KN-LIO) systems that leverage online SDF decoding and iterated error-state Kalman filtering to fuse laser and inertial data. Our KN-LIO minimizes information loss and improves accuracy in state estimation, while also accommodating asynchronous multi-LiDAR inputs. Evaluations on diverse high-dynamic datasets demonstrate that our KN-LIO achieves performance on par with or superior to existing state-of-the-art solutions in pose estimation and offers improved dense mapping accuracy over pure LiDAR-based methods. The relevant code and datasets will be made available at https://**.

Problem

• Lidar-IMU Fusion
• Detailed Mapping
• Neural Fields

Innovation

• KN-LIO
• Neural Fields
• Real-time Localization and Mapping