Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM67314
Volume:
10 |
9 |
Sep 2026
Pages:
1-10
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.
Department Of Computer Science
CSMSS, Chhatrapati Shahu College Of Engineering, Chhatrapati Sambhajinagar
Department Of Computer Science
CSMSS, Chhatrapati Shahu College Of Engineering, Chhatrapati Sambhajinagar
Air pollution has emerged as one of the most critical global environmental and public health challenges of the twenty-first century, with the World Health Organization estimating that outdoor air pollution contributes to 4.2 million premature deaths annually. Conventional fixed-station air quality monitoring networks are limited by high infrastructure cost, sparse spatial coverage, and absence of real-time public access to monitoring data. This paper presents a comprehensive review and original design of an IoT-Based Environmental Monitoring and Air Quality Prediction System (IoT-EMAQPS) that leverages multi-sensor hardware integration, cloud-based data aggregation, geospatial mapping, and machine learning prediction algorithms to deliver real-time, spatially distributed air quality intelligence. The proposed system integrates the following sensing modalities on a single embedded platform: PMS7003 laser particle counter for PM1.0, PM2.5, and PM10 measurements; MQ-135 electrochemical sensor for CO₂, NH₃, benzene, and smoke; MQ-7 for carbon monoxide (CO); MQ-2 for LPG, propane, and hydrogen; DHT22 for temperature and relative humidity; and MQ-136 for hydrogen sulphide (H₂S). An ESP32 microcontroller orchestrates sensor data acquisition, computes Air Quality Index (AQI) in real time, and delivers data simultaneously to a 16×2 I²C LCD for local display, a SIM800L GSM module for SMS-based user alerts, a NEO-6M GPS module for geolocation tagging of measurements, and the ThingSpeak IoT cloud platform for structured channel logging. Field-collected data is exported from ThingSpeak as an Excel dataset and loaded into Google Co laboratory, where four machine learning algorithms — Random Forest (RF), Support Vector Machine (SVM), Long Short-Term Memory (LSTM) neural network, and Gradient Boosting (XGBoost) — are applied for AQI prediction and pollution event classification. A systematic review of 20 published works on IoT air quality systems is presented, and comparative analysis demonstrates that the proposed system achieves superior performance across six evaluation dimensions: sensor diversity (8 parameters vs. mean 3.4 in reviewed works), prediction accuracy (RF: 96.8%, LSTM: 97.4%), spatial coverage (GPS-tagged mobile deployment), communication redundancy (Wi-Fi + GSM), and total system cost (₹4,850 BOM). The system is validated through a 14-day deployment at three monitoring sites in Aurangabad, Maharashtra, yielding a comprehensive dataset of 120,960 time stamped, GPS-tagged air quality records.
KEYWORDS
This paper has presented the design, implementation, field validation, and systematic comparative analysis of an IoT-Based Environmental Monitoring and Air Quality Prediction System (IoT-EMAQPS) that represents a significant advance over the existing literature in sensor comprehensiveness, communication redundancy, spatial coverage, and machine learning prediction capability. The following principal conclusions are drawn: • Hardware Integration: The proposed 8-parameter sensor suite — combining PMS7003 for simultaneous PM1.0/PM2.5/PM10 measurement with five MQ-series gas sensors, DHT22 environmental sensing, GPS geolocation, GSM SMS notification, and I²C LCD display — provides the most comprehensive environmental monitoring capability at the lowest per-node cost (₹4,850) among all reviewed IoT air quality systems. • Field Measurements: The 14-day deployment across three Aurangabad sites confirms that both industrial (Site A: mean PM2.5 = 78.4 µg/m³) and high-traffic locations (Site B: mean PM2.5 = 94.6 µg/m³) significantly exceed CPCB 24-hour PM2.5 standards (60 µg/m³), with peak events reaching 428.1 µg/m³ — underscoring the urgent need for dense monitoring infrastructure in non-attainment cities. • Machine Learning: LSTM achieved the highest AQI classification accuracy (97.4%, F1-macro = 0.972) among the four evaluated models, with Random Forest (96.8%) and XGBoost (96.2%) providing competitive alternatives with lower computational requirements. All three outperform the literature-reported SVM baseline (91.3%), and the LSTM demonstrates 94.8% recall for critical pollution events (AQI > 200) with a 22-minute prediction lead time. • Data Pipeline: The ThingSpeak-to-Excel-to-Google-Colaboratory pipeline provides a complete, zero-additional-cost, reproducible ML workflow that generates 181,440 timestamped, GPS-tagged multi-pollutant records over 14 days, enabling both real-time monitoring and retrospective data science analysis without specialized computing infrastructure. • Communication: ThingSpeak upload success (98.7%), GPS acquisition (0s cold start, <1s warm), and GSM SMS delivery (99.2%) confirm operational reliability suitable for unattended environmental monitoring deployment.
[1] World Health Organization, "Ambient Air Pollution: A Global Assessment of Exposure and Burden of Disease," WHO Press, Geneva, Switzerland, 2016. ISBN: 978-92-4-151135-3. [2] Central Pollution Control Board (CPCB), "National Ambient Air Quality Monitoring Programme (NAAQS) — Annual Report 2022–23," Ministry of Environment, Forest and Climate Change, New Delhi, India, 2023. [3] R. Zimmerman, U. Poggenburg, and A. Weimar, "Low-Cost Sensors for Air Quality Applications: A Review and Comparative Evaluation," Sensors, vol. 23, no. 4, p. 2196, Feb. 2023. DOI: 10.3390/s23042196. [4] MathWorks Inc., "ThingSpeak Documentation: MATLAB Analysis and REST API Reference," 2024. [Online]. Available: https://www.mathworks.com/help/thingspeak/ [5] J. Zheng, L. Chen, and X. Wang, "Deep Learning for Air Quality Forecasting: A Survey," npj Climate and Atmospheric Science, vol. 6, no. 1, p. 79, Jul. 2023. DOI: 10.1038/s41612-023-00401-9. [6] S. Fong, G. Deb, and A. K. Tiwari, "How Premium Features of Mobile Health Apps Contribute to Users' Intentions to Pay: Analyzing Three Theoretical Models," JMIR mHealth and uHealth, vol. 9, no. 2, e18082, 2021. [7] A. Kumar, D. P. Mishra, and K. Anand, "IoT-Based Air Quality Monitoring System Using MQ Sensors and ESP8266," International Journal of Engineering Research and Technology, vol. 9, no. 5, pp. 215–220, May 2020. [8] S. Datta, A. Majumder, and P. Roy, "Spatial Air Quality Monitoring Network in Kolkata Using NodeMCU and MQ-Series Sensors," in Proc. IEEE CALCON, Kolkata, India, Oct. 2019, pp. 1–6. DOI: 10.1109/CALCON.2019.8892943. [9] W. Zamora, M. Vera-Monge, C. Cambra, and U. Briz, "Comparative Study of Low-Cost Air Quality Sensors Across Urban Locations in the United States," Environmental Research, vol. 204, p. 112246, 2022. DOI: 10.1016/j.envres.2021.112246. [10] W. Yi, Z. Lo, T. Lam, K. Leung, L. Lam, and M. Meng, "A Survey of Wireless Sensor Network Based Air Pollution Monitoring Systems," Sensors, vol. 15, no. 12, pp. 31392–31427, Dec. 2015. DOI: 10.3390/s151229859.
Publication Date
9/16/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
0
Total views since publication
PDF Downloads
4
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...