The monitoring and evaluation of road traffic noise requires a reliable road traffic noise prediction method.The study conducts the affective artificial neural network(EANN)method to construct the traffic noise of some streets in a city,uses two different scenarios and different input combinations,compares the EANN model with the classical feedforward neural network(FFNN),and verifies the validity of the EANN model.The results show that when the EANN model is applied to road traffic noise prediction,the prediction efficiency of FFNN and empirical model is improved by 14％and 37％respectively in the verification stage.By classifying traffic into subclasses(in scenario 1)and feeding it into the model,the performance of EANN and FFNN models increases by 8％and 12％in the validation phase,respectively.The sensitivity analysis of input parameters shows that the total traffic volume is the most important factor affecting the road traffic noise in the study area,followed by the number of cars,the number of medium vehicles,the number of heavy vehicles,the average speed and the proportion of heavy vehicles.
维普
万方数据