In-pipe Robot Perception for Challenging Altered Environments

Gunatilake, Delpachchitra Arachchige Amal

In-pipe Robot Perception for Challenging Altered Environments

Gunatilake, Delpachchitra Arachchige Amal

Permalink

Publication Type:: Thesis
Issue Date:: 2021

Open Access

Copyright Clearance Process

Recently Added
In Progress
Open Access

This item is open access.

Adobe PDF

Download contents and abstractAdobe PDF (357.77 kB)

Adobe PDF

Download thesisAdobe PDF (33.4 MB)

View statistics

Full metadata record

Field	Value	Language
dc.contributor.author	Gunatilake, Delpachchitra Arachchige Amal
dc.date.accessioned	2022-08-28T22:36:19Z
dc.date.available	2022-08-28T22:36:19Z
dc.date.issued	2021
dc.identifier.uri	http://hdl.handle.net/10453/160974
dc.description	University of Technology Sydney. Faculty of Engineering and Information Technology.	en_US.UTF-8
dc.description.abstract	Robotics can play a crucial role in the condition assessment of critical infrastructure assets such as underground drinking water pipes. Currently, water utilities worldwide spend billions of dollars every year to reliably inspect and rehabilitate corroding and deteriorating pipes. Internal pipe linings are widely used as a renewal method to increase structural strength. Post pipe-lining quality assurance and long-term performance monitoring of the applied liners are essential for maintaining pipe assets. In this regard, this thesis focuses on the development of a multi-sensor approach to liner defect mapping in underground human-altered environments. A mobile robotic sensing system that can scan, detect, locate, and measure internal pipeline defects is proposed. This is achieved by generating three-dimensional RGB-D maps using stereo camera vision in combination with an infrared laser profiling unit. The system does not involve complex calibration procedures and utilises orientation correction to provide accurate real-time RGB-D maps. Defects are identified and colour mapped for easier visualisation. The robotic sensing system was extensively tested under laboratory conditions, followed by field deployment in buried water pipes in Sydney, Australia. The experimental results showed that the RGB-D maps were generated with millimetre-level accuracy and with demonstrated liner defect quantification. The accuracy of the map is dependent on the robot localisation. Therefore, a cost-effective UHF-RFID tags were used for robot localisation inside pipelines. The results showed that unlike outdoor RFID localisation, inside the pipeline, the signal behaves uniquely, which makes the localisation task challenging and unique. Signal processing using a Gaussian process combined particle filter was applied to accurately localise the robot. Experiments carried out on field-extracted pipe samples from the Sydney Water pipe network showed that using the RSSI and Phase data together in the measurement model with the particle filter algorithm improves the localisation accuracy up to millimetre-level, through utilisation of a two-antenna sensor model. Robot localisation assumes an accurate map. In pipes, this is tedious and therefore SLAM is desirable. A novel solution for SLAM using UHF-RFID signal processing for underground pipe environments is proposed. The problem was formulated as a Graph-SLAM combining signal cross-correlation and mapping with respect to the RFID sensor measurements. Experiments in the laboratory showed that the solution can localise the robot with 2.5-centimetres accuracy while building the RFID map. The results showed that the solution allows accurate identification of defect locations in a 50-meter long pipe, and performs vastly better than standard encoder-based localisation methods.	en_US.UTF-8
dc.format	Thesis (PhD)
dc.language.iso	en_US	en_US.UTF-8
dc.relation	https://opus.lib.uts.edu.au/bitstream/10453/160974/2/02whole.pdf
dc.rights	The author owns the copyright in this thesis including all reproduction and reuse rights for the work. The work may not be altered without the permission of the copyright owner. Attribution is essential when quoting or paraphrasing from this thesis.
dc.rights	au.edu.uts.lib/ppc
dc.rights	info:eu-repo/semantics/openAccess
dc.title	In-pipe Robot Perception for Challenging Altered Environments	en_US.UTF-8
dc.type	Thesis
utslib.copyright.status	open_access	*

Abstract:

Robotics can play a crucial role in the condition assessment of critical infrastructure assets such as underground drinking water pipes. Currently, water utilities worldwide spend billions of dollars every year to reliably inspect and rehabilitate corroding and deteriorating pipes. Internal pipe linings are widely used as a renewal method to increase structural strength. Post pipe-lining quality assurance and long-term performance monitoring of the applied liners are essential for maintaining pipe assets. In this regard, this thesis focuses on the development of a multi-sensor approach to liner defect mapping in underground human-altered environments. A mobile robotic sensing system that can scan, detect, locate, and measure internal pipeline defects is proposed. This is achieved by generating three-dimensional RGB-D maps using stereo camera vision in combination with an infrared laser profiling unit. The system does not involve complex calibration procedures and utilises orientation correction to provide accurate real-time RGB-D maps. Defects are identified and colour mapped for easier visualisation. The robotic sensing system was extensively tested under laboratory conditions, followed by field deployment in buried water pipes in Sydney, Australia. The experimental results showed that the RGB-D maps were generated with millimetre-level accuracy and with demonstrated liner defect quantification. The accuracy of the map is dependent on the robot localisation. Therefore, a cost-effective UHF-RFID tags were used for robot localisation inside pipelines. The results showed that unlike outdoor RFID localisation, inside the pipeline, the signal behaves uniquely, which makes the localisation task challenging and unique. Signal processing using a Gaussian process combined particle filter was applied to accurately localise the robot. Experiments carried out on field-extracted pipe samples from the Sydney Water pipe network showed that using the RSSI and Phase data together in the measurement model with the particle filter algorithm improves the localisation accuracy up to millimetre-level, through utilisation of a two-antenna sensor model. Robot localisation assumes an accurate map. In pipes, this is tedious and therefore SLAM is desirable. A novel solution for SLAM using UHF-RFID signal processing for underground pipe environments is proposed. The problem was formulated as a Graph-SLAM combining signal cross-correlation and mapping with respect to the RFID sensor measurements. Experiments in the laboratory showed that the solution can localise the robot with 2.5-centimetres accuracy while building the RFID map. The results showed that the solution allows accurate identification of defect locations in a 50-meter long pipe, and performs vastly better than standard encoder-based localisation methods.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/160974