Enhanced Adaptive Monte Carlo Localization with Iterative Closest Point Scan Matching for Agricultural Mobile Robots
International Journal of Advanced Computer Science and Applications · 2026 · 17 (10)
DOI: 10.14569/ijacsa.2026.01710104 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Advanced Computer Science and Applications, 2026. Verfasst von Brian Lai Lap Hong und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Accurate localization of agricultural mobile robots is difficult in plantations because repeated rows of crops or trees provide limited distinctive geometry. Previously, an Adaptive Monte Carlo Localization (AMCL) refined by the Normal Distributions Transform (NDT) was proposed for two-dimensional Light Detection and Ranging (LiDAR) localization to improve the localization accuracy of the robots. This paper extends that study by cascading Iterative Closest Point (ICP) scan matching after AMCL on every cycle the current AMCL estimate initializes point-to-point scan-to-map registration, allowing ICP to refine the pose using local tree-trunk geometry, and the refined pose is published as the output of the pipeline. The method was evaluated in a Gazebo palm oil plantation with ROS 2 Humble, a Clearpath Husky platform, and a 270◦ 2D LiDAR with a 30 m maximum range. Conventional AMCL, AMCL with NDT, and AMCL with ICP were evaluated against Gazebo ground truth using Absolute Pose Error, reported both after rigid-body trajectory alignment and in the map frame without alignment, and rotational and translational relative pose error. Compared with conventional AMCL, the proposed method reduced mean aligned absolute pose error from 0.413 m to 0.305 m and its root mean square error from 0.427 m to 0.323 m. Mean rotational relative pose error decreased from 1.118 degrees to 0.113 degrees, while its root mean square error decreased from 1.465 degrees to 0.223 degrees. The ICP-based extension also achieved lower translational and rotational errors than the NDT-based method, which was run from the same node with the same inputs and preprocessing. The results indicate that ICP refinement of the AMCL estimate is effective in repetitive plantation geometry, although isolated scan-matching failures remain.
Zitieren
APA
Hong, B. L. L., Izhar, M. A. M., & Ahmad, N. (2026). Enhanced Adaptive Monte Carlo Localization with Iterative Closest Point Scan Matching for Agricultural Mobile Robots. International Journal of Advanced Computer Science and Applications, 17(10). https://doi.org/10.14569/ijacsa.2026.01710104
Vancouver
Hong BLL, Izhar MAM, Ahmad N. Enhanced Adaptive Monte Carlo Localization with Iterative Closest Point Scan Matching for Agricultural Mobile Robots. International Journal of Advanced Computer Science and Applications. 2026;17(10). doi:10.14569/ijacsa.2026.01710104
BibTeX
@article{Hong2026,
author = {Hong, Brian Lai Lap and Izhar, Mohd Azri Mohd and Ahmad, Norulhusna},
title = {Enhanced Adaptive Monte Carlo Localization with Iterative Closest Point Scan Matching for Agricultural Mobile Robots},
journal = {International Journal of Advanced Computer Science and Applications},
year = {2026},
volume = {17},
number = {10},
doi = {10.14569/ijacsa.2026.01710104}
}
EUVIMED