AI-Enabled IoT-Based Real-Time Air Quality and Comfort Monitoring System

This paper has presented an AI-enabled, IoT-based prototype for real-time indoor air quality and comfort monitoring, designed specifically for educational institutions. The system successfully addresses the three principal gaps identified in the literature and market survey: it introduces a novel holistic metric (CACI), it operates proactively rather than reactively through cloud-based time-series forecasting, and it achieves this at a cost accessible to public educational institutions. Key outcomes include: (i) a functional hardware prototype validated in both simulation and physical deployment; (ii) a CACI algorithm that unifies thermal comfort and air quality into a single, actionable score; (iii) seamless ThingSpeak cloud integration supporting real-time data logging, six-channel visualisation, and MATLAB-based CACI forecasting; and (iv) a demonstrated proactive alert capability providing approximately 12 minutes of advance warning before WARNING-level conditions are reached. Future work will focus on replacing the MQ-135 proxy with a calibrated NDIR CO2 sensor (e.g., MH-Z19) for higher measurement fidelity, expanding the sensor suite to include TVOC detection, deploying a TensorFlow Lite LSTM model directly on the ESP32-S3 for edge-based prediction independent of cloud connectivity, and developing a dedicated mobile application for facility managers. Ultimately, this work serves as a compelling proof of concept for scalable, intelligent environmental monitoring in resource-constrained institutional settings.

[1] J. Saini, M. Dutta, and G. Marques, "Indoor Air Quality Monitoring Systems Based on Internet of Things: A Systematic Review," Int. J. Environ. Res. Public Health, vol. 17, no. 14, p. 4942, Jul. 2025. [2] M. S. Kanitkar and R. Sahu, "Indoor Air Quality Prediction Using Machine Learning Techniques," Int. J. of Sci. and Res. (IJSR), vol. 11, no. 6, pp. 123–127, Jun. 2024. [3] M. Santamouris et al., "Indoor Environmental Quality (IEQ) and Student Performance: A Review," Energy and Buildings, vol. 179, pp. 323–341, 2018. [4] S. T. Singh et al., "A review of machine learning applications in time-series prediction for environmental data," J. Environ. Manage., vol. 182, pp. 500–512, 2021. [5] A. Kumar et al., "A Low-Cost Air Quality Monitoring System using IoT and Machine Learning for Smart Cities," J. Ambient Intell. Humaniz. Comput., 2021. [6] Espressif Systems, "ESP32-WROOM-32E Datasheet," Ver. 3.3, 2024. [Online]. Available: https://www.espressif.com [7] Aosong, "AM2302 (DHT22) Digital Temperature and Humidity Sensor Datasheet," 2016. [8] Z. Li et al., "Predicting PM2.5 Concentrations Using LSTM and GRU Models," IEEE Access, vol. 8, pp. 104762–104771, 2020. [9] D. Snyder et al., "Calibration of Low-Cost Air Quality Sensors using Machine Learning," Atmos. Meas. Tech., vol.13, pp. 1797–1820, 2020. [10] F. J. Garcia et al., "An IoT Architecture for Real-Time CO2 Monitoring in University Classrooms," Sensors, vol. 19, no. 2, p. 441, 2019. [11] MathWorks, "ThingSpeak IoT Platform Documentation," 2025. [Online]. Available: https://www.mathworks.com/help/thingspeak/ [12] A. K. Singh, IoT Based Projects, BPB Publications, 2023.