An Asthma Occupant Focused Digital Twin for Indoor Air Quality in Residential Environments

Khosh Amadi, Negin (2026) An Asthma Occupant Focused Digital Twin for Indoor Air Quality in Residential Environments. Doctoral thesis, Birmingham City University.

Text Negin Khosh Amadi PhD Thesis_Final Version_Final Award March 2026.pdf - Accepted Version Restricted to Repository staff only until 30 April 2028. Download (4MB)

Abstract

Indoor air quality (IAQ) has a critical influence on respiratory health, particularly for individuals with asthma who are highly sensitive to short-term pollutant exposure. Despite advances in IAQ sensing technologies, existing approaches typically operate as isolated monitoring tools with limited behavioural context, minimal predictive capability, and little capacity for personalised decision support. These limitations constrain their effectiveness in real residential settings, where pollutant fluctuations are strongly shaped by occupant activities, ventilation patterns, and spatial conditions.

To address these gaps, a Digital Twin (DT) framework, termed the Air Quality and Health-Responsive Digital Twin (AIR-DT), was developed to support asthma-sensitive IAQ management. Unlike existing IAQ monitoring platforms, AIR-DT integrates multi-parameter real-time monitoring, spatially contextualised visualisation, adaptive short-term prediction, historical data analysis, and personalised intervention logic within a unified architecture. The framework shifts IAQ management from reactive threshold-based alerts to proactive, context-aware guidance tailored to vulnerable occupants. Through IoT and AI capabilities, AIR-DT predicts pollutant trends and delivers actionable insights to help residents maintain healthier indoor conditions before asthma-relevant thresholds are exceeded.

Design Science Research (DSR) was adopted to structure the development and evaluation of the AIR-DT artefact. Functional requirements were established through a synthesis of evidence on asthma-relevant pollutants, behaviour–environment interactions, and limitations of existing IAQ solutions. These requirements informed the AIR-DT architecture and its integrated sensing, modelling, and intervention workflows.

AIR-DT was implemented and evaluated across two residential case studies. IoT sensors captured IAQ parameters (PM₂.₅, PM₁₀, CO₂, TVOCs, temperature, humidity) and recorded window-opening events as the behavioural indicator, forming the data foundation for real-time analytics. Multiple machine learning and deep learning models were assessed for forecasting, with ensemble-based approaches yielding the highest accuracy in short-term pollutant prediction. The intervention engine translated environmental and behavioural insights into personalised recommendations aligned with asthma-sensitive thresholds, while the user-centred interface presented real-time, historical, and forecasted IAQ data in an accessible and actionable form.

The AIR-DT system was evaluated using the Framework for Evaluation of Digital Systems (FEDS), conceptual, analytical, descriptive, and naturalistic assessment, supported by empirical performance testing, expert interviews, and occupant feedback. Case Study 1 enabled validation and refinement of sensing workflows, modelling pipelines, and intervention rules, while Case Study 2 demonstrated the applicability, robustness, and scalability of the refined system in a different residential context. Across both implementations, AIR-DT proved capable of capturing multi-source data reliably, predicting pollutant dynamics with appropriate accuracy, and supporting proactive IAQ management through personalised, context-sensitive guidance.

This research contributes to theory by (1) advancing understanding of asthma-relevant IAQ requirements through the integration of environmental, behavioural, and spatial determinants;(2) defining the core architectural components of a residential, health-oriented Digital Twin;(3) embedding occupant–IAQ activity relationships into DT logic to support personalised intervention design; (4) demonstrating the role and comparative performance of short-term IAQ forecasting models; (5) introducing a model for proactive, personalised intervention; and (6) illustrating the use of the FEDS framework for evaluating real-time environmental DTs.

Contributions to practice include (1) a replicable workflow for integrating multi-parameter sensing with behavioural data; (2) a practical implementation of ML-based IAQ forecasting for proactive household management; (3) a personalised intervention engine tailored to real behaviours and room-level conditions; (4) a user-centred visualisation interface; (5) an in-depth analysis layer that supports long-term decision-making; (6) a modular, scalable DT architecture suitable for adoption across diverse residential settings.

Overall, the study demonstrates that DT technology can provide an effective, user-centred, and scalable framework for improving indoor environmental health in asthma-sensitive households and establishes a foundation for future work on automated intervention, expanded behavioural sensing, and integration with broader building performance systems.

Item Type:	Thesis (Doctoral)
Dates:	Date Event 23 March 2026 Accepted
Uncontrolled Keywords:	Digital Twin, Visualisation, Monitoring, Intervention, Indoor air quality, Asthma, Respiratory disease
Subjects:	CAH13 - architecture, building and planning > CAH13-01 - architecture, building and planning > CAH13-01-02 - building
Divisions:	Architecture, Built Environment, Computing and Engineering > Architecture and Built Environment > Built Environment Doctoral Research College > Doctoral Theses Collection
Depositing User:	Louise Muldowney
Date Deposited:	14 Apr 2026 11:56
Last Modified:	14 Apr 2026 11:56
URI:	https://www.open-access.bcu.ac.uk/id/eprint/16970

Actions (login required)

View Item