Assessment and mlr modeling of traffic noise at major urban roads of residential and commercial areas of surat city

Abstract

In a developing country like India, traffic noise pollution is becoming a severe problem due to inefficient public transport in urban areas. The number of personal vehicles is increasing due to the lack of convenient public transportation. The study of urban road traffic noise is an important issue. Traffic noise prediction models are playing a very important role in the decision making and proper implementation of rules and regulations for a certain area. This research work was carried out to develop a traffic noise prediction model for two rigid and one flexible pavement road of Surat city using multiple linear regression. Significant factors affecting traffic noise such as classified traffic volume count, road width, and average building height were selected as input parameters for a detailed survey. Models have been developed using the data of three roads separately and one final model has also been developed using the data of all three roads. According to the data obtained from the detailed survey, it is found that the minimum equivalent traffic noise level is exceeding the permissible limits on all roads. Among the prediction in three urban roads, the predicted output result from the MLR model showed poor correlation with an average absolute % error of 2.065 and an R2 value of 0.25. But with the combined road there is a slight improvement in the statistical values, viz. average absolute % error 2.27 and R2 = 0.51. This also proves that dependent variable equivalent noise (Leq) is not linearly dependent on independent variables, classified traffic volume count, road width, and average building height. To overcome this problem, road traffic noise prediction models may be developed using evolutionary computing tools like genetic algorithm, neural networks, etc.