Autonomous localization and navigation for a railway inspection and repair system

Simple item page
dc.contributor.advisorDurazo Cardenas, Isidro S.
dc.contributor.advisorStarr, Andrew
dc.contributor.authorRahimi, Masoumeh
dc.date.accessioned2025-06-25T08:45:35Z
dc.date.available2025-06-25T08:45:35Z
dc.date.freetoread2025-06-25
dc.date.issued2023-04
dc.descriptionStarr, Andrew - Associate Supervisor
dc.description.abstractRobotic and autonomous systems have brought numerous benefits to various industries, such as increased accuracy, safety, efficiency, cost-effectiveness, and reduced time. The railway industry has also leveraged robotic technologies for track maintenance jobs, albeit their application is often restricted to specific use cases. This research introduces the Robotic Inspection and Repair System (RIRS), which operates both on and around the railway track. A key consideration for autonomous systems is the need for absolute localization, which is essential for maintenance systems on the railway track. Therefore, this thesis focuses on implementing and developing an autonomous localization and navigation system for the RIRS, using the Global Positioning System (GPS). However, due to the railway environment complexity which includes electromagnetic interferences, tunnels, and dense vegetation, GPS inevitably degrades, making vehicle localization extremely challenging. The RIRS localization system is investigated in two separate modes: off-track and on-track. For the off-track phase, to achieve a higher frequency rate and accurate robot pose estimation even in GPS-denied environments, the Extended Kalman Filter (EKF) filter is applied to fuse continuous data with global pose estimates. This approach's effectiveness is also compared with the Real-Time Appearance-based Mapping (RTAB-Map) approach's odometry based on absolute and relative pose error. For the on-track phase, the RIRS aimed to identify track defects at the absolute level initially, but this is infeasible due to GPS unavailability in almost 20% of the railway network. Therefore, first, the RIRS starts navigating autonomously using GPS odometry. Then track-side object detection and 3D pose estimation is implemented to compensate for the error caused by GPS. The average error of 0.07 m in the vehicle's location demonstrates the reliability of this strategy for a maintenance vehicle.
dc.description.coursenamePhD in Manufacturing
dc.identifier.urihttps://dspace.lib.cranfield.ac.uk/handle/1826/24078
dc.language.isoen
dc.publisherCranfield University
dc.publisher.departmentSATM
dc.rights© Cranfield University, 2023. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright holder.
dc.subjectROS
dc.subjectSensor Fusion
dc.subjectExtended Kalman Filter
dc.subjectObject detection
dc.subject3D Pose Estimation
dc.subjecton track
dc.subjectoff track
dc.titleAutonomous localization and navigation for a railway inspection and repair system
dc.typeThesis
dc.type.qualificationlevelDoctoral
dc.type.qualificationnamePhD

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Rahimi_M_2023.pdf
Size:
7.42 MB
Format:
Adobe Portable Document Format
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.63 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

PhD, EngD and MSc by research theses (SATM)