Neural network modelling of vehicle fuel consumption and NOx emissions based on data clustering

Abstract

Since automobiles are becoming increasingly popular, fuel consumption has become a crucial factor to consider when purchasing a vehicle. Car emissions, one of the causes of global warming, have become a big worry as the planet warms. As a result, the fuel efficiency and exhaust emissions of automobiles are tracked in this article using neural networks based on DBSCAN clustering. This study uses DBSCAN to cluster and model NOx emissions and vehicle fuel consumption. Using DBSCAN, the vehicle speed, acceleration, fuel use, and NOx emission data are first grouped and pre-processed into many data clusters. Then, each data cluster develops a corresponding feedforward neural network model to simulate NOx emission. The experimental results show that the method suggested in this study substantially improves the model’s dynamic reaction time and modeling accuracy.