Dependable monitoring with low-cost sensor networks and visualisation for urban air quality

Le, Hoang Trung

Dependable monitoring with low-cost sensor networks and visualisation for urban air quality

Le, Hoang Trung

Permalink

Publication Type:: Thesis
Issue Date:: 2024

Open Access

Copyright Clearance Process

Recently Added
In Progress
Open Access

This item is open access.

Adobe PDF

Download thesisAdobe PDF (21.61 MB)

View statistics

Full metadata record

Field	Value	Language
dc.contributor.author	Le, Hoang Trung
dc.date.accessioned	2025-12-18T23:08:43Z
dc.date.available	2025-12-18T23:08:43Z
dc.date.issued	2024
dc.identifier.uri	http://hdl.handle.net/10453/190995
dc.description	University of Technology Sydney. Faculty of Engineering and Information Technology.	en_US.UTF-8
dc.description.abstract	Health risks posed by a polluted atmosphere have significantly escalated in severity. Therefore, the importance of accessibility to reliable air quality monitoring and forecasting has been widely emphasised in recent years. In response, wireless sensor networks are deployed as complementary data sources to regulatory monitoring stations. However, ensuring the reliability and accuracy of air quality sensor data for public dissemination and forecasting models requires robust solutions. This thesis introduces a fault-tolerant sensing system realised by employing a dependable monitoring configuration for redundancy. Furthermore, the system adopts Dempster-Shafer theory, which effectively handles uncertain and conflicting evidence, for validating sensor readings against reference-grade stations. The application of this theory enables the identification of the most reliable sensor within a network of redundant low-cost sensors, enhancing data reliability prior to training predictive models. The resulting monitoring configuration demonstrates improved robustness and consistency in environmental sensing, particularly under fluctuating air quality conditions. Thereafter, the validated data is streamlined to an ensemble deep learning nowcasting framework, which dynamically selects the best-performing model based on changing environmental conditions. By leveraging the Dempster-Shafer theory for data fusion, the adaptive framework utilises the strengths of multiple models to enhance the robustness of air quality nowcasting. As for real-world impact, this work demonstrates the operation of a visualisation dashboard, bridging the gap between real-time monitoring and deep learning-based predictions for an urban air quality monitoring network. The proposed system aims to improve early detection and decision-making with a user-friendly interface and comprehensible format, ensuring that communities are better prepared for air quality challenges. All efforts reiterate the primary objectives of this research, which are to enhance public health protection and support sustainable environmental initiatives.	en_US.UTF-8
dc.format	Thesis (ME)
dc.language.iso	en_US	en_US.UTF-8
dc.relation	https://opus.lib.uts.edu.au/bitstream/10453/190995/1/thesis.pdf
dc.rights	info:eu-repo/semantics/openAccess
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	© 2025 Hoang Trung Le
dc.rights	au.edu.uts.lib/cph
dc.title	Dependable monitoring with low-cost sensor networks and visualisation for urban air quality	en_US.UTF-8
dc.type	Thesis
utslib.copyright.status	open_access	*

Abstract:

Health risks posed by a polluted atmosphere have significantly escalated in severity. Therefore, the importance of accessibility to reliable air quality monitoring and forecasting has been widely emphasised in recent years. In response, wireless sensor networks are deployed as complementary data sources to regulatory monitoring stations. However, ensuring the reliability and accuracy of air quality sensor data for public dissemination and forecasting models requires robust solutions. This thesis introduces a fault-tolerant sensing system realised by employing a dependable monitoring configuration for redundancy. Furthermore, the system adopts Dempster-Shafer theory, which effectively handles uncertain and conflicting evidence, for validating sensor readings against reference-grade stations. The application of this theory enables the identification of the most reliable sensor within a network of redundant low-cost sensors, enhancing data reliability prior to training predictive models. The resulting monitoring configuration demonstrates improved robustness and consistency in environmental sensing, particularly under fluctuating air quality conditions. Thereafter, the validated data is streamlined to an ensemble deep learning nowcasting framework, which dynamically selects the best-performing model based on changing environmental conditions. By leveraging the Dempster-Shafer theory for data fusion, the adaptive framework utilises the strengths of multiple models to enhance the robustness of air quality nowcasting. As for real-world impact, this work demonstrates the operation of a visualisation dashboard, bridging the gap between real-time monitoring and deep learning-based predictions for an urban air quality monitoring network. The proposed system aims to improve early detection and decision-making with a user-friendly interface and comprehensible format, ensuring that communities are better prepared for air quality challenges. All efforts reiterate the primary objectives of this research, which are to enhance public health protection and support sustainable environmental initiatives.

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

http://hdl.handle.net/10453/190995