Published paper

LiDAR-Inertial SLAM for Low-Cost Ground Robots in GPS-Denied Warehouses (TEST)

A A. Hassan M M. Abdelrahman S S. El-Sayed N N. Farouk

Published in

IEEE ICRA

Year

2026

DOI

10.1109/ICRA.2026.1043277

Publisher page DOI Code

Abstract

Warehouse automation increasingly depends on robots that can localise reliably without GPS and without the cost of survey-grade sensors. We present a tightly-coupled LiDAR-inertial odometry pipeline targeting sub-$500 sensor stacks, combining a sparse voxel map with an IMU pre-integration factor graph. On a 12-hour continuous run across three warehouse sites we achieve 0.38% relative translational drift, a 41% improvement over the strongest open-source baseline at comparable compute, while sustaining 20 Hz on an ARM Cortex-A76.

Why this matters

Most warehouse robots on the market today either rely on floor markers or on sensor suites that cost more than the robot itself. Our goal with this work was to see how far a commodity 16-beam LiDAR and a $12 IMU can be pushed.

Approach

The pipeline has three stages:

Pre-integration of IMU measurements between LiDAR sweeps
Sparse voxel matching against a rolling local map
Factor-graph optimisation over a sliding window of keyframes

The key insight is that in structured indoor environments, aggressive voxel sparsification costs almost nothing in accuracy but buys a 3x speedup.

Results

Method	Drift	Rate
Baseline A	0.64%	10 Hz
Baseline B	0.71%	14 Hz
Ours	0.38%	20 Hz

Full source and the warehouse dataset are available in the linked repository.

Figures

Cite this paper

@inproceedings{hassan2026lislam,
  title     = {LiDAR-Inertial SLAM for Low-Cost Ground Robots in GPS-Denied Warehouses},
  author    = {Hassan, A. and Abdelrahman, M. and El-Sayed, S. and Farouk, N.},
  booktitle = {IEEE International Conference on Robotics and Automation (ICRA)},
  year      = {2026},
  doi       = {10.1109/ICRA.2026.1043277}
}

Details

Venue: IEEE ICRA
Year: 2026
Added: Jul 13, 2026
Topics: researchrobotics

Interested in collaborating on research with our team?

Get in touch

All publications