Toward Prediction of Roadside PM2.5 Concentration: A Multi-Factor Prediction Method

Predicting particulate matter 2.5 (PM2.5) concentration in exhaust from traffic emissions is helpful for environmental protection and also helps to protect human health. Previous studies on the prediction of PM2.5 mostly use meteorological condition models on a macroscopic level, which ignore the influence of traffic emissions on a microscopic level. We propose predicting the concentration of roadside PM2.5 on a microscopic level by using a multi-factor method that combines the comprehensive air quality model (CAMx) with the traffic flow condition model and wavelet neural network. This study concluded that 40% of the concentration of PM2.5 comes from traffic emissions by using the CAMx model, and the accuracy of the multi-factor prediction method was 86.17%. This paper builds a multi-factor prediction method to predict the concentration of roadside PM2.5, which has practical significance for reducing the concentration of PM2.5 emissions under the microscopic level.