AI-BASED AIR QUALITY MONITORING

Air pollution is a significant environmental and public-health concern, while conventional monitoring stations can be costlyand may not provide dense, street-level coverage. This paper proposes an AI-based air quality monitoring framework thatcombines low-cost sensing, Internet of Things (IoT) communication, data validation, and machine-learning-assisted trendanalysis. A sensor node collects particulate-matter and selected environmental measurements, an edge controller checks dataquality, and a web dashboard presents time-stamped observations, trends, and alerts. The design separates direct sensormeasurements from estimated Air Quality Index (AQI) categories and model-based forecasts, preventing predictions frombeing presented as official measurements. A rule-based baseline is proposed first; machine-learning models such as RandomForest and time-series models can be evaluated when sufficient labelled historical data are available. The paper describesarchitecture, data handling, security, evaluation metrics, limitations, and a practical implementation plan. It is a designproposal and does not claim field-tested accuracy or measured pollution reduction. Calibration against reference instrumentsand validation under local conditions are required before operational use.