The Development of a Real-Time Air Quality Monitoring and PM2.5 Alert System Using Internet of Things Technology

Abstract

This research aimed to develop a real-time air quality monitoring and PM2.5 alert system using Internet of Things technology, evaluate the system’s performance, and assess user acceptance. The study followed a research-and-development approach. The developed system consisted of three main components: an outdoor air quality monitoring station, a cloud-based data processing and storage system, and a display and user notification module. The monitoring station was powered by solar energy with backup batteries and was capable of continuous outdoor operation. It measured PM2.5 and temperature and transmitted the data through an IoT gateway over a 4G network to the cloud system using the MQTT protocol. The data were displayed through a web dashboard and an on-site LED display, while notifications were sent via Telegram when PM2.5 levels exceeded the specified thresholds. System performance was evaluated using black-box testing and error assessment. All ten test items passed, covering measurement, data transmission, processing, display, notification, risk-level classification, and overall system operation. Measurement agreement was evaluated using 720 time-aligned paired hourly observations collected over 30 days from the AQMS and an IQAir AirVisual commercial monitor. The comparison yielded a Mean Absolute Percentage Error of 7.72%, a Mean Absolute Error of 1.65 µg/m³, and a Root Mean Squared Error of 1.96 µg/m³. These results indicated agreement between the measurements from the two devices under the study location, period, and environmental conditions. However, the comparison did not confirm the system’s metrological accuracy because it was not calibrated against an FRM/FEM standard instrument. The user acceptance evaluation was conducted with 30 participants, and overall acceptance was at the highest level, with a mean of 4.51 and a standard deviation of 0.51. The findings indicate that the system performed the predefined functions and produced measurements consistent with those of the comparison device under the conditions of this study. However, calibration against a standard instrument, long-term testing, and evaluation with a larger user sample should be conducted before wider deployment.