DeepPointMap: Advancing LiDAR SLAM with Unified Neural Descriptors
DOI:
https://doi.org/10.1609/aaai.v38i9.28909Keywords:
ROB: Localization, Mapping, and Navigation, CV: 3D Computer Vision, ML: Deep Learning Algorithms, ML: Representation LearningAbstract
Point clouds have shown significant potential in various domains, including Simultaneous Localization and Mapping (SLAM). However, existing approaches either rely on dense point clouds to achieve high localization accuracy or use generalized descriptors to reduce map size. Unfortunately, these two aspects seem to conflict with each other. To address this limitation, we propose an unified architecture, DeepPointMap, achieving excellent preference on both aspects. We utilize neural network to extract highly representative and sparse neural descriptors from point clouds, enabling memory-efficient map representation and accurate multi-scale localization tasks (e.g., odometry and loop-closure). Moreover, we showcase the versatility of our framework by extending it to more challenging multi-agent collaborative SLAM. The promising results obtained in these scenarios further emphasize the effectiveness and potential of our approach.Downloads
Published
2024-03-24
How to Cite
Zhang, X., Ding, Z., Jing, Q., Zhang, Y., Ding, W., & Feng, R. (2024). DeepPointMap: Advancing LiDAR SLAM with Unified Neural Descriptors. Proceedings of the AAAI Conference on Artificial Intelligence, 38(9), 10413-10421. https://doi.org/10.1609/aaai.v38i9.28909
Issue
Section
Intelligent Robots (ROB)