Computational driver behavior models for vehicle safety applications
Doctoral thesis, 2023

The aim of this thesis is to investigate how human driving behaviors can be formally described in mathematical models intended for online personalization of advanced driver assistance systems (ADAS) or offline virtual safety evaluations. Both longitudinal (braking) and lateral (steering) behaviors in routine driving and emergencies are addressed. Special attention is paid to driver glance behavior in critical situations and the role of peripheral vision.

First, a hybrid framework based on autoregressive models with exogenous input (ARX-models) is employed to predict and classify driver control in real time. Two models are suggested, one targeting steering behavior and the other longitudinal control behavior. Although the predictive performance is unsatisfactory, both models can distinguish between different driving styles.

Moreover, a basic model for drivers' brake initiation and modulation in critical longitudinal situations (specifically for rear-end conflicts) is constructed. The model is based on a conceptual framework of noisy evidence accumulation and predictive processing. Several model extensions related to gaze behavior are also proposed and successfully fitted to real-world crashes and near-crashes. The influence of gaze direction is further explored in a driving simulator study, showing glance response times to be independent of the glance's visual eccentricity, while brake response times increase for larger gaze angles, as does the rate of missed target detections.

Finally, the potential of a set of metrics to quantify subjectively perceived risk in lane departure situations to explain drivers' recovery steering maneuvers was investigated. The most influential factors were the relative yaw angle and splay angle error at steering initiation. Surprisingly, it was observed that drivers often initiated the recovery steering maneuver while looking off-road.

To sum up, the proposed models in this thesis facilitate the development of personalized ADASs and contribute to trustworthy virtual evaluations of current, future, and conceptual safety systems. The insights and ideas contribute to an enhanced, human-centric system development, verification, and validation process. In the long term, this will likely lead to improved vehicle safety and a reduced number of severe injuries and fatalities in traffic.

Driver adaptation

Driver models

Visual attention

Safety benefit assessment

Evidence accumulatoin

PrARX

Hybrid dynamical systems

ADAS

Predictive processing

Room Omega in the Jupiter building, Hörselgången 5
Opponent: Professor Brett Fajen, Rensselaer Polytechnic Institute, USA

Author

Malin Svärd

Chalmers, Mechanics and Maritime Sciences (M2), Vehicle Safety

Online driver behavior classification using probabilistic ARX models

16th International IEEE Conference on Intelligent Transportation Systems, October 2013, The Hague, The Netherlands,;(2013)p. 1107-1112

Paper in proceeding

A study of appropriate model complexity for estimation of car-following behavior

Proceedings of the 3rd International Symposium on Future Active Safety Technology Towards zero traffic accidents, 2015,;(2015)p. 137-144

Paper in proceeding

A quantitative driver model of pre-crash brake onset and control

Proceedings of the Human Factors and Ergonomics Society,;Vol. 61(2017)p. 339 - 343

Paper in proceeding

Svärd, M., Bärgman, J., Markkula, G., & Ljung Aust, M., Using naturalistic and driving simulator data to model driver responses to unintentional lane departures

More than a million lives are lost on the roads each year, and driver behavior is one of the main contributing factors in most of today's vehicle crashes. Modern cars are equipped with advanced driver assistance systems (ADAS) to prevent crashes, or at least make them less severe. These systems are designed to warn the driver of upcoming dangerous situations, and sometimes also brake or steer.

Since ADASs interact with the driver, their effectiveness may be improved by adapting the warnings and interventions to individual preferences, using models of driver behavior. Driver models can also be used to evaluate the ADASs' overall impact on road safety to better understand the human-system-vehicle interaction and its consequences. An affordable way to do this is through computer simulations with a virtual vehicle and driver.

This thesis demonstrates how computational driver models can be constructed based on data from driving simulator studies and real-world driving. The proposed models are simple enough to represent an intuitive understanding of driver behavior, and most of them strive to mimic the human mind. Since gaze direction and peripheral vision influence drivers' responses in critical situations, these factors are also investigated.

To sum up, driver models enable computer simulations with realistic driver reactions and have the potential to improve ADASs. Consequently, in the long term, the amount of killed or severely injured people on the roads should decrease.

Improved quantitative driver behavior models and safety assessment methods for ADAS and AD (QUADRIS)

VINNOVA (2020-05156), 2021-04-01 -- 2024-03-31.

Quantitative Driver Behaviour Modelling for Active Safety Assessment Expansion (QUADRAE)

VINNOVA (2015-04863), 2016-01-01 -- 2019-12-31.

Subject Categories

Other Computer and Information Science

Other Engineering and Technologies not elsewhere specified

Applied Psychology

Other Mathematics

Vehicle Engineering

Robotics

Computer Science

Computer Vision and Robotics (Autonomous Systems)

Areas of Advance

Information and Communication Technology

Transport

Health Engineering

ISBN

978-91-7905-840-1

Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 5306

Publisher

Chalmers

Room Omega in the Jupiter building, Hörselgången 5

Online

Opponent: Professor Brett Fajen, Rensselaer Polytechnic Institute, USA

More information

Latest update

5/9/2023 2