Vehicle Performance Assessment Using the OBD2 Port and Artificial Neural Network

There is a growing body of literature that recog-nises the importance of vehicle performance assessment to evaluate and improve vehicle dynamics and fuel consumption. This study set out to investigate the usefulness of the artificial neural networks (ANNs) to predict the vehicle performance curves and acceleration responses. The experimental measurements are obtained from an OBD2 port of a Suzuki SX4 sedan, and the torque-power engine curves are achieved according to the SAE standard SAEJ1491 JUL2006. Therefore, it is demonstrated that the vehicle performance can be improved and predicted using vehicle measurements, gear ratios, and dynamic rolling radius. Then, the actual paper reports that ANNs can be em-ployed as a non-parametric model to predict vehicle behaviour, improve comfort, and reduce the steps between gear changes.