Application of Artificial Neural Network to Predict Exhaust Emissions from Road Transport

Abstract

Vehicle manufacturers have to meet the standards due to the emission standard limitations. For this reason, every new vehicle must satisfy the limits of emission regulations by passing the driving cycle test accepted by their markets. However, even the new cars satisfy emission limits, our environment is being polluted more than expectations due to the old vehicles used in transportations and unrepresentative driving cycle for real world conditions so that real time exhaust emissions were analyzed in this study. It is aimed to calculate the exhaust gas released by road transports by using IPCC second approach method in Istanbul. Then an artificial neural network model was developed to predict a correlation between real-time exhaust emissions and vehicle number, mean speed. With using different training functions, it is demanded to define the optimum percentage error between the target and the predicted values. It was observed that the ANN model can predict exhaust gases with correlation coefficient in the range of 0.97–0.99. This study shows that the created ANN can be used to accurately predict the greenhouse gas in Istanbul.

Keywords: Artificial neural network, greenhouse gas, IPCC second approach