Regression analysis for selection of a noise descriptor to accurately predict traffic noise on Indian highways

The need for controlling community noise is being felt with increasing intensity each passing day. Studies have shown that motor vehicles operating on high-ways consti-tute one of the major sources of noise, which have an impact on the community. Various prediction models have been developed the world over, but need has always been felt of a model, which could be used to accurately predict noise on Indian highways. In order to achieve this, it is essential to use a suitable descriptor for predicting
traffic noise. This present study was conducted for identi-fying a noise descriptor which is accurate in predicting noise under Indian Conditions. An extensive literature
survey revealed that L„, and Leg have been mostly used worldwide to describe highway noise. This study focused on collection and analysis of data to establish which one of the above descriptors was suitable for use in the devel-opment of a noise prediction model. The literature survey further revealed that the procedure outlined in the CRTN model (Department of Transport and Welsh Office. Calcu-lation of Road Traffic Noise) was best suited for data collection, analysis and noise prediction. The CRTN
model uses Li„ as the noise descriptor. Analysis of the data indicated that the error of prediction for L10 was quite less than that for that Leg. Further analysis esta-blished that Li() was a suitable descriptor, as compared to L , for predicting noise on Indian Highways.