APPLICATION OF ARTIFICIAL NEURAL NETWORK IN TRAFFIC NOISE POLLUTION MODELLING

Abstract

This research has been motivated by the fact that present road traffic noise prediction models have not been improved since their development in the 1970s and 1980s, although road traffic noise nuisance is a significant and growing issue in India and elsewhere as the number of vehicles on roads is increasing day by day. Noise is a global problem due to several factors: the increase in the number of per capita vehicles, the increase in demographic density, appliances and vehicles capable of generating loud noise, and also the fact that society is getting used to higher noise levels. Since 1972, when the WHO classified noise as a pollutant, most of the industrialized countries have decided to regulate noise through laws or local regulations. In this study Artificial Neural Network (ANN) has been applied to predict noise pollution level in Delhi, capital city of India. Factors that predominantly influence noise pollution level in a traffic noise model framework were classified into two categories: traffic volume and traffic speed. Traffic volume, traffic speed and noise level data of traffic were collected at six identified locations in the city. The structure of the model selected consisted of input variable as 2W, 3W, Car, Jeep, Van, Bus, Truck, and Leq as output variable and corresponding traffic speed on both sides of the road were taken as input data and Leq as output variable. Models based on forward-propagation neural network were trained, validated and tested using the data collected through field studies. All other data collected were the position of the measurement station, the geographical situation between the noise source and the measurement station, wind speed and direction, air temperature and relative humidity, and time of day. The model was selected by varying the number of hidden neurons from 3 to 10. The best model was selected on the basis of Mean Square Error (MSE), which was in present case of 3 hidden neurons. The model selected can be applied for the prediction of Leq level. It is shown that the artificial neural networks can be a useful tool for the prediction of noise with sufficient accuracy in the observed time intervals.