FS-DICP: A Fast and Simple Point-to-Point Iterative Closest Point Algorithm With Doppler Velocity
IEEE Signal Processing Letters · 2026 · 33 · 2974-2978
DOI: 10.1109/lsp.2026.3711527 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Zhicong Xie und 4 weiteren Autor:innen. Die Arbeit nennt 21 Literaturangaben.
Abstract (Originalsprache)
With the rapid advancement of frequency-modulated continuous-wave (FMCW) LiDAR technology, four-dimensional (4D) point clouds have garnered significant attention in the fields of autonomous driving and robotic navigation. The 4D point cloud encapsulates the 3D geometric information of individual points, along with a scalar representing Doppler velocity as the fourth dimension—an extension that effectively enhances the representational capacity of dynamic scenes. In this work, we propose a streamlined 3-degree-of-freedom (3-DoF) Iterative Closest Point (ICP) algorithm that leverages Doppler velocity measurements to estimate ego-motion while filtering dynamic points. By modeling the intrinsic relationships among velocity, rotation, and translation, we develop a 3-DoF ICP framework that decouples rotational and translational components for efficient computation. Experiments on two public datasets (one synthetic and one real-world dataset) demonstrate that our method outperforms competing approaches in both real-time registration and the translation component of relative pose error (RPE Trans).
Zitieren
APA
Xie, Z., Zhu, W., Wang, D., Huang, H., & Yang, G. (2026). FS-DICP: A Fast and Simple Point-to-Point Iterative Closest Point Algorithm With Doppler Velocity. IEEE Signal Processing Letters, 33, 2974-2978. https://doi.org/10.1109/lsp.2026.3711527
Vancouver
Xie Z, Zhu W, Wang D, Huang H, Yang G. FS-DICP: A Fast and Simple Point-to-Point Iterative Closest Point Algorithm With Doppler Velocity. IEEE Signal Processing Letters. 2026;33:2974-2978. doi:10.1109/lsp.2026.3711527
BibTeX
@article{Xie2026,
author = {Xie, Zhicong and Zhu, WeiXuan and Wang, Di and Huang, Haozhi and Yang, Guanghua},
title = {FS-DICP: A Fast and Simple Point-to-Point Iterative Closest Point Algorithm With Doppler Velocity},
journal = {IEEE Signal Processing Letters},
year = {2026},
volume = {33},
pages = {2974--2978},
doi = {10.1109/lsp.2026.3711527}
}
EUVIMED