Designing of KNN-based Air Quality Predicting System

Air is the basis for human survival. PM2.5 is an important indicator of air quality. Predicting PM2.5 is an important means to ensure good air quality. In order to better predict PM2.5, an environmental monitoring system based on photovoltaic power supply and STM32 control is designed. The monitoring system is powered by a large-capacity lithium battery, and solar photovoltaic panels are used to charge the lithium battery; GP2Y1014AU is a PM2.5 data sensor that collects data for the data acquisition module; STM32 serves as the controller for the control module. The assembled monitoring system is deployed outdoors to collect PM2.5 data and uploaded to the Ali Cloud platform through the Message Queuing Telemetry Transport (MQTT) protocol. In order to predict PM2.5, a dataset is constructed using the collected data, and the sliding window and K-Nearest Neighbor (KNN) methods are employed to make short-term predictions of air quality. The experimental results show that the KNN model performs well in predicting PM2.5. Compared with the sliding window model, its Mean Absolute Error (MAE), Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) are reduced by 82%, 69% and 81% respectively.