In English

Improved driver model for testing in HIL environment

Harald Freij
Göteborg : Chalmers tekniska högskola, 2013. Diploma work - Department of Applied Mechanics, Chalmers University of Technology, Göteborg, Sweden, ISSN 1652-8557; 2013:76, 2013.
[Examensarbete på avancerad nivå]

This master's thesis concerns driver models, with focus on non perfect driving. On the base of a black box driver behaviour model, perturbations are added for modelling lane drifting, non-alert driving (for example caused by sleepiness) and collision from behind. The driver models are developed and implemented for usage in an HIL environment. This means that both high and low level behaviour is considered. The running off road is modelled as a small steering wheel angle offset introduced with a smooth ramp. This has support in data from extensive field operation tests. The non alert driving is modelled both as driving with random input update frequency, and a sinusoidal steering wheel angle offset. These models are qualitatively compared to data corresponding to sleepy driving, recognised by a Driver Alert Control safety function. The behaviour connected to collisions from behind is modelled using Gipps' longitudinal model, with small presumed decelerations. The thesis also presents different low level model approaches, that is how a target vehicle motion is realised, in terms of pedal positions and steering wheel angle. Both PID and open loop methods are discussed. It is shown that PID controlling can give a very good resemblance between target and obtained motion, but also that a high correlation can be acquired without a good resemblance on the low level input.

Nyckelord: Driver model, HIL, Testing environments, Run off road, Non alert driving, Collision from behind, Field operation test

Publikationen registrerades 2014-01-08. Den ändrades senast 2014-02-11

CPL ID: 191772

Detta är en tjänst från Chalmers bibliotek