The road transport has become the major source of environmental degradation in urban centres. It produces negative externalities (i.e. pollution, delay, etc.) that are usually connected with the queues of traffic flows and the congestion of the road network.The quantitative estimation of roadside pollutant levels is very complex. Many variables are involved such as the type of vehicle (characterized by different antipollution devices, used fuels, engine temperatures, maintenance level of engines, etc.), the different cinematic conditions of the flows, the urban/road network structure, the weather conditions, etc.Therefore it is important to develop scientific tools able to predict roadside pollutant levels, in order to apply the best strategy to improve air quality and to obtain atmospheric pollution levels within the regulatory limits. So traffic management measures can help public administrations to reach their objectives in the field of traffic pollutant reduction in the urban centres.An initial exploration of the scientific literature revealed as the neural networks can be a useful tool in the field of the traffic pollution modelling. They are able to capture very complex and non linear correlations among the variables involved in the problem.The aim of our study is the calibration of a neural network to forecast the roadside pollutant levels, using the data measured by the Automatic Urban Network in different points of the urban transport network of Palermo, in order to evaluate different traffic management measures able to reduce traffic pollutants, such as NOx and CO.
|Number of pages||13|
|Publication status||Published - 2015|