Abstract :
The present paper illustrates study on feasibility of ANN modeling for road traffic noise prediction at Indian Intermediate, Yavatmal city, district place of Vidarbha region in Maharashtra state. Sixteen locations were identified at uninterrupted and interrupted traffic flow conditions for conducting field studies Traffic volume study (composition & classified traffic volume) and noise level study are carried out simultaneously. Artificial neural network software (Elite ANN) is used, the network uses feed forward negative back propagation algorithm with three hidden and three previous time elements of weights (Pandharipande & Badhe, 2002). ANN modeling is performed through input data as- Total traffic, Traffic composition (bus/truck, LCV, TW, bicycle and others) in % and carriageway width, Distance of receiver from pavement. Output is estimated as L10, Leq, LNP, TNI and NC. The observed input and output data is processed and trained through ANN for interrupted and uninterrupted flow condition. To enhance the accuracy of prediction, further this model has been tested by using linear regression analysis between observed and predicted noise levels. In the present case, it is observed that there is no significant difference between the observed and predicted noise levels, indicating the accuracy of model.
Keywords :
backpropagation; environmental science computing; feedforward neural nets; noise pollution; regression analysis; road traffic; artificial neural network software; feedforward negative backpropagation algorithm; linear regression analysis; road traffic noise descriptor; uninterrupted traffic flow condition; Artificial neural networks; Bicycles; Cities and towns; Feeds; Noise level; Predictive models; Roads; Software algorithms; Telecommunication traffic; Traffic control; ANN (Artificial Neural Network); Noise levels; Road Traffic;
