WIT Press

Prediction Of Carbon Monoxide Concentration Near Roads By Means Of Artificial Neural Networks


Free (open access)

Paper DOI








399 kb


W. Kaminsky & E. Tomczak


Prediction of carbon monoxide concentration near roads by means of artificial neural networks W. Kaminsky & E. Tomczak Faculty of Process and Environmental Engineering, Technical University of Lodz, Poland Abstract The artificial neural, networks (ANN) as a tool to predict air pollution were presented taking into account meteorological conditions and parameters which characterise the source of pollutants. A comparison was made between the two methods for calculation of carbon monoxide concentration in the region of a city road. The first method was based on a hybrid model which was a combination of ANN (a neural model based on radial basis functions -RBF) and the Pasquille model. In the other method the multilayer perceptron -MLP only, was applied to predict the level of carbon monoxide near the roadside edge. Topologies and the flow diagrams of signals in both networks were given and statistical estimation of the two methods was presented. 1 Introduction Along narrow roads in built-up areas (cities), the state of atmosphere affected by pollutants emitted by vehicles is dramatic. This phenomenon is particularly intensive when in spring, autumn or winter cars with not heated enough engines drive on the roads. The cars may be equipped with autocatalysts system but it