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

Purpose and goal:

For efficient development of advanced driver support systems and self-driving cars, it is necessary to be able to assess their potential safety benefit during development. QUADRIS wants to improve three parts of the simulation-based process used for such assessments. First, we want to develop methods that ensure that the collisions simulated are representative.

Second, we want to develop a validated driver model for run-off-road accidents. Third, we want to expand the driver modeling so that it also includes self-driving cars in the form of so-called reference-driver models. Expected results and effects:

Having the capacity to perform safe, fast and accurate safety benefit assessments for ADAS and AD during development will be essential for the Swedish automotive industry to stay competitive; and even more so when automated driving becomes a key determinant of brand success and competitiveness. Approach and implementation:

The project will run for four years and the work is centered around the three doctoral students who will carry out most of the work. This includes both data collection, building simulation structures and doing modeling to the extent needed to meet the goals within each work package.