A new study has been published investigating social driving behaviors and self-driving cars.
The SWOV Institute for Road Safety Research has issued the report ‘Social driving behavior in automated driving’ which highlights a gap in current automated driving system (ADS) regulations— how they account for social driving behavior.
SWOV researchers said that while human drivers regularly engage in socially aware driving—like yielding to others or anticipating unpredictable moves—ADS are not yet required or tested to do the same.
The team found that current testing methods—such as virtual simulations, real-world trials, and in-service monitoring—are not well suited to detect social driving flaws in ADS.
Key interactions with human road users may go unnoticed, especially since many systems rely on after-the-fact reporting or narrow testing conditions, the study said.
The SWOV Institute for Road Safety Research said it could have implications for the safety of these vehicles: if an automated car lacks pro-social behaviors, it might inadvertently block traffic, create confusion, or even cause crashes.
The authors have argued for clearer definitions, better testing tools such as advanced driving simulators with better models of human behavior, and early identification of social interaction issues.
“Without a proactive, top-down approach to regulating social driving in ADS, safety improvements may come only after incidents occur—defeating the very purpose of automation,” the SWOV Institute for Road Safety Research said.
