Citation

Galarza MA, Paradells J. Int. J. Veh. Safety 2018; 10(2): 138-161.

Copyright

(Copyright © 2018, Inderscience Publishers)

DOI

10.1504/IJVS.2018.094176

PMID

unavailable

Abstract

The increasing amount of infotainment services available in vehicles makes it necessary to devise a system capable of managing how information should be delivered and accessed in accordance with the driving complexity scenario. The objective of this study is to provide a useful model for categorising driving scenarios in terms of their complexity. For this purpose, data collected from driving tests are analysed employing data mining techniques and machine learning methods for finding the more influential variables of driving complexity. The input variables used are associated with primary driving tasks and road characteristics available in current vehicles. As a result, the most relevant variables that enable the categorisation of the driving scenario are identified and a model capable of predicting driving complexity in real time is constructed. Given the model accuracy obtained, a practical application could be the adaptation of Human Machine Interfaces (HMI).

Keywords: driving complexity; primary driving task; data mining; machine learning; vehicle safety; human machine interface.

Language: en