Event

As we look to the future of driver behavior models, key questions arise: What challenges must be tackled? What opportunities and risks do these models present? And which topics should we prioritize within the SAFER platform? On September 18th, we will explore these questions at SAFER's event on Driver Behaviour Models. Featuring insights from projects like QUADRIS and V4SAFETY, this event will provide a platform for in-depth discussions on advancing driver behavior research. Join us to contribute your organisation’s perspectives and help shape the future direction within the SAFER community.

What a Driver behaviour model is

A driver behavior model is a computational or theoretical representation that simulates and predicts how a driver behaves in various driving situations. These models typically encompass aspects such as decision-making processes, reaction times, attention, and the driver’s responses to environmental factors like road conditions, traffic, and vehicle dynamics. Driver behavior models are used in a wide range of applications, including the development of Advanced Driver Assistance Systems (ADAS), autonomous vehicles, traffic simulations, and safety assessments. They help researchers and engineers understand and anticipate how human drivers might interact with technology and their environment, ultimately contributing to the design of safer and more efficient transportation systems.

The event is part of the SAFER Research Day, se full agenda here.
 

AGENDA

09:00 Welcome and today's agenda

09:10 Introduction to Driver Behaviour models

09:15 Key Insights from Improved Quantitative Driver Behavior Models and Safety Assessment Methods for ADAS and AD (QUADRIS), Malin Svärd, Volvo Cars 

09:30 Are the UNECE ALKS reference driver models really competent and careful (QUADRIS)?, Pierluigi Olleja, Chalmers University

09:40 Human behavior modeling in simulation-based pre-crash safety assessment – opportunities and challenges (the V4SAFETY project and beyond), Jonas Bärgman, Chalmers University (and preliminary one PhD student that have been working in the QUADRIS project focusing on reference models for automated driving)

09:50 Discussion session: The future of driver models in safety assessments

In this discussion, we will explore the future of driver models, a rapidly emerging area gaining significance in vehicle safety assessments. As driver models become part of rating protocols, we'll explore their benefits and potential challenges. These models improve safety ratings by accurately simulating driver behavior in various traffic situations. We will also discuss the collaborative research needed and the role that SAFER's multidisciplinary research environment can play in supporting stakeholders to create value through these advancements.

10:20 Networking fika and poster session

Welcome!

More about the projects

Improved Quantitative Driver Behavior Models and Safety Assessment Methods for ADAS and AD (QUADRIS)
QUADRIS is following the projects QUADRA and QUADRAE. This four-year project, running from April 1, 2021, to March 31, 2025, aims to enhance the development of advanced driver support systems and self-driving cars by refining safety assessment methods.

The project focuses on three main areas: developing methods to ensure simulated collisions are representative, creating validated driver models for run-off-road accidents, and expanding driver modeling to include self-driving cars through reference-driver models. These improvements will enable safer, faster, and more accurate safety benefit assessments during the development of ADAS and AD systems. QUADRIS will play a crucial role in maintaining the competitiveness of the Swedish automotive industry, especially as automated driving becomes a key factor in brand success.

V4Safety
The V4SAFETY method aims to enhance road safety by providing a reliable framework for evaluating various safety measures within Connected, Cooperative, and Automated Mobility (CCAM). By integrating driver models in virtual safety assessment, this framework will allow for a more detailed analysis of how different human behaviors and safety measures, including in-vehicle technologies and infrastructure changes, impact traffic safety. However, the inclusion of computational driver models does not ensure that the evaluation process can accurately simulate driving, crashing and the safety impact of new safety measures. That said, there are large benefits with behavior-model based virtual safety assessment, if it is done correctly and transparently. Simulation-based pre-crash safety assessment has the potential to save lives – helping us to move towards Vision Zero – by supporting the decision-making for a variety of stakeholders, from system developers, via policy makers, to consumers.