Modeling and Simulation of Road Traffic Noise Using Artificial Neural Network and Regression.

Area of Science:

• Environmental Science
• Urban Planning
• Data Science

Background:

• Urban noise pollution poses significant environmental and health challenges.
• Accurate prediction of traffic noise is crucial for effective urban planning and mitigation strategies.
• Existing noise modeling techniques may require enhancement for complex urban environments.

Purpose of the Study:

• To develop and compare artificial neural network (ANN) and regression models for predicting in-city road traffic noise pollution.
• To assess the suitability of these models using noise measurements and vehicle count data.
• To identify the more accurate prediction model for urban noise levels.

Main Methods:

• Data collection included noise measurements and vehicle counts at three urban locations over 12 hours.
• Two predictive models were developed: artificial neural network and regression.
• Simulations were performed using MATLAB and DATAFIT software.
• Model performance was evaluated by comparing predicted noise levels with measured data using statistical metrics.

Main Results:

• Both artificial neural network and regression models demonstrated suitable performance in predicting traffic noise pollution.
• The artificial neural network model exhibited predictions closer to the experimentally measured noise levels.
• Statistical metrics such as normalized bias and root mean squared error indicated the superiority of the ANN model.

Conclusions:

• Artificial neural networks offer a promising approach for accurately modeling and predicting urban traffic noise pollution.
• The developed ANN model can be a valuable tool for urban planners and environmental agencies.
• Further research can explore incorporating additional variables to enhance noise prediction accuracy.