Date of Award
2025-08-01
Degree Name
Master of Science
Department
Manufacturing Engineering
Advisor(s)
Amit J. Lopes
Abstract
Indoor air quality (IAQ) is a critical public health concern, especially in enclosed environments where people spend extended periods. This thesis presents a cost-effective, real-time IAQ monitoring and control system that integrates IoT hardware with supervised machine learning (ML) to enable intelligent assessment and automated response. The system incorporates BME688, PMSA003I, SGP30, SGP40, and MQ-135 sensors connected to a Raspberry Pi 5 via I²C and SPI protocols. A Python-based pipeline collects, logs, and visualizes environmental data using a Dash web interface. Both regression and classification tasks were performed to model indoor air quality indicators. Regression models were developed to predict PM2.5, TVOC, and eCOâ?? concentrations using six different algorithms. Among these, Random Forest yielded the best performance achieving the highest R² score for all targets. For classification, air quality was categorized into three levels such as Good, Moderate, and Unhealthy using a percentile-based thresholding method. Six classification models were evaluated using metrics such as accuracy, F1-score, and ROC AUC, with SMOTE applied to mitigate class imbalance. XGBoost demonstrated the best overall classification performance, achieving approximately 80% accuracy, strong generalization across classes, and the highest ROC AUC (>0.90). Consequently, XGBoost was selected as the final model for deployment. The trained model was integrated into a real-time inference on the Raspberry Pi, which activates GPIO-based relay control in response to predicted air quality conditions. This work demonstrates a scalable, intelligent IAQ solution suitable for edge deployment in smart homes and buildings. Future directions include time-series modeling, smart HVAC integration, and multi-zone cloud-based analytics.
Language
en
Provenance
Received from ProQuest
Copyright Date
2025-08
File Size
113 p.
File Format
application/pdf
Rights Holder
Md Abubakar Siddique
Recommended Citation
Siddique, Md Abubakar, "IoT-Enabled Indoor Air Quality Monitoring and Control Using Machine Learning Techniques" (2025). Open Access Theses & Dissertations. 4476.
https://scholarworks.utep.edu/open_etd/4476