LiDAR-Based Positioning with 3-D Digital Maps: Comparative Analysis and Multi-Solid-State LiDAR Integration
Loading...
Authors
Elsayed, Mohamed Mokhtar Abdelaziz Hussein
Date
2025-05-23
Type
thesis
Language
eng
Keyword
LiDAR , Positioning , Perception sensors
Alternative Title
Abstract
Accurate and reliable positioning is fundamental for the safe and effective navigation of autonomous vehicles (AVs). While the global navigation satellite system (GNSS) traditionally serves as the primary positioning solution, it becomes unreliable or unavailable in dense urban environments and indoor scenarios due to its limitations, such as signal obstruction and multipath interference. Consequently, light detection and ranging (LiDAR)-based navigation approaches, particularly those relying on LiDAR-to-map registration, have emerged as promising alternatives. However, these methods depend critically on sensor characteristics and the robustness of registration algorithms.
This thesis investigates the impact of LiDAR technology choice and registration algorithm selection on positioning performance in AV applications. Initially, a comparative study is conducted to characterize the differences between mechanical spinning LiDAR(MSL) and solid-state LiDAR(SSL). Real-world data are used to examine critical sensor attributes, including scanning patterns, field of view (FOV), and resulting point cloud density, highlighting how these factors influence map registration outcomes. Further, a modular LiDAR-to-map registration (LMR) pipeline is developed and utilized to systematically evaluate the iterative closest point (ICP) and normal distribution transform (NDT) registration algorithms across diverse driving conditions. The results indicate distinct advantages and limitations for each sensor-algorithm combination, emphasizing the necessity for tailored solutions in varying operational contexts.
Building upon these insights, this thesis introduces a decentralized multi-SSL fusion framework aimed at overcoming inherent SSL limitations such as limited FOV and susceptibility to occlusions or degraded map environments. The proposed architecture integrates pose corrections from multiple independently operating SSLs through a novel adaptive weighting strategy, considering registration quality metrics, spatial deviations from inertial measurements, and sensor reliability. Extensive experimental validation in real urban scenarios demonstrates that this decentralized approach improves positioning accuracy, fault tolerance, and overall robustness against common AV navigation challenges, including partial sensor failures and environmental uncertainties.
The contributions presented here offer practical insights into designing and deploying resilient, scalable, and accurate LiDAR-based navigation systems, thereby advancing AV navigation capabilities, especially in GNSS-challenged scenarios.
Description
Citation
Publisher
License
Queen's University's Thesis/Dissertation Non-Exclusive License for Deposit to QSpace and Library and Archives Canada
Intellectual Property Guidelines at Queen's University
Copying and Preserving Your Thesis
This publication is made available by the authority of the copyright owner solely for the purpose of private study and research and may not be copied or reproduced except as permitted by the copyright laws without written authority from the copyright owner.
Attribution-NonCommercial-ShareAlike 4.0 International
Intellectual Property Guidelines at Queen's University
Copying and Preserving Your Thesis
This publication is made available by the authority of the copyright owner solely for the purpose of private study and research and may not be copied or reproduced except as permitted by the copyright laws without written authority from the copyright owner.
Attribution-NonCommercial-ShareAlike 4.0 International