Title: Multi-input multi-output intelligent modelling techniques and application to human driver

Authors: Emre Tekin; Seniz Ertugrul

Addresses: Graduate School of Science, Engineering and Technology, Istanbul Technical University, 34496 Maslak, Istanbul, Turkey ' Mechanical Engineering Department, Istanbul Technical University, 34437 Beyoglu, Istanbul, Turkey

Abstract: Human driver models to achieve longitudinal and lateral control of vehicle are developed without complex mathematical modelling techniques. For this purpose, intelligent modelling techniques are reviewed and application of these techniques to driver modelling problem is studied. Among these techniques, the ones which show satisfactory performance are selected to model human driver and the obtained driver models are tested in virtual environment. Data to train driver model are obtained by a real human driver in virtual driving environment to achieve longitudinal and lateral driving actions. A road profile that is different from the training road profile is used to validate the human driver model.

Keywords: intelligent modelling; human driver models; fuzzy logic; artificial neural networks; ANNs; driving simulation; multi-input multi-output; MIMO modelling; virtual human drivers; human-machine interaction; HMI; vehicle control; mathematical modelling; lateral control; longitudinal control.

DOI: 10.1504/IJVP.2016.079921

International Journal of Vehicle Performance, 2016 Vol.2 No.4, pp.390 - 417

Accepted: 30 Aug 2015
Published online: 19 Oct 2016 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article