In English

Driver Behaviour Model Based Threat Assessment for Forward Collision Warning Systems

Abhishek Karunagaran
Göteborg : Chalmers tekniska högskola, 2018. Master's thesis - Department of Mechanics and Maritime Sciences; 2018:42, 2018.
[Examensarbete på avancerad nivå]

Forward Collision Warning systems warn the driver when there’s a risk of a collision with a car or truck in front of the equipped vehicle. Different drivers have different driving styles, and an attempt is made to come up with a conceptual design of such a system that adapts to these differences. The aim of this thesis was to develop a threat assessment algorithm that estimates the driver’s “comfort zone” by continuously analyzing vehicle signals, and uses it to decide when to issue a forward collision warning to the driver. A literature survey of relevant driver behaviour models for braking was performed for this application, and estimation schemes were designed and developed for a looming threshold based, and an evidence accumulation based model. Further, a test track study was conducted to collect driving data, and the developed estimators were tested on this data. Qualitative comparisons of the two driver behaviour models were made, and used to propose conceptual designs for threat assessment algorithms. Due to the design of the test track study which used professional test drivers, and involved repetitive tasks, the data collected was not suitable to draw conclusions on the performance of the developed estimators. A comparison of the obtained estimates of driver model parameters, and parameter values reported in literature showed potential but this needs to be verified with a larger naturatlistic driving dataset.

Nyckelord: forward collision warning, driver behaviour models, braking behaviour, driver adaptation, evidence accumulation

Publikationen registrerades 2018-10-01. Den ändrades senast 2018-10-01

CPL ID: 256055

Detta är en tjänst från Chalmers bibliotek