News

A car can already warn, brake or steer to help a driver in a dangerous situation. But when should it step in? That sounds like a simple question, but it really isn’t. Imagine that an obstacle suddenly appears in front of the car. Technically, the car may still be able to make a sharp turn and avoid it. But can an ordinary driver spot the danger, understand what is happening and steer away quickly enough?

This is the question at the heart of the Be-Point pre-study. The project looks at what researchers call a behavioural point of no return: the point where we can no longer reasonably expect the driver to avoid a crash because of human physical or cognitive limitations.

The car may be capable of more than the driver
Many safety systems today are largely based on what the vehicle itself is technically capable of doing. But people don’t react instantly. We need time to notice a danger, understand the situation and do something about it. Stress and surprise can also affect how well we drive in that moment. So, there may be a gap between what the car can theoretically do and what a person can actually manage.

Take a car approaching a stationary vehicle. A theoretical driver with an instant reaction might still be able to steer around it at the very last moment. For a real person, that opportunity may already be gone. If the safety system waits too long, its help may come too late. If it has a better understanding of human capabilities, it could warn, brake or assist with steering at a better point in time.

Be-Point will investigate whether this gap can be described and measured in a useful way.

First, we need to know what can actually be measured
Be-Point is a pre-study, so the aim is not to develop a finished safety system. The researchers will review existing research and data to find useful ways of measuring human performance, including how much and how quickly a driver can steer in a critical situation. The project will also look at whether a driver’s reliance on automation can be described as something measurable, and how models of human behaviour could be included in existing simulation tools.

The longer-term idea is fairly straightforward: future safety systems should not only be based on “What can the car do?”, but also on “What can we reasonably expect the driver to do?”

That could help safety systems intervene at the right time, making them both more effective and easier for drivers to understand and trust. The expected road safety benefit is better crash avoidance and injury reduction by basing interventions on observed human capabilities, rather than vehicle capabilities alone.

A first step towards better human-centred safety systems
If the pre-study shows that this can be done, the results can provide a basis for a larger research project. In the future, these kinds of models could also be used when testing new driver assistance and automated driving systems in simulation. Instead of only asking whether the vehicle can handle a situation, we could also test whether the system gives a real driver a reasonable chance to handle it.

Through SAFER, the results can also be discussed and shared with a wider network of researchers, companies and other road safety actors. This gives the work a better chance of being picked up, challenged and taken further in future research and development.

At its core, Be-Point is trying to answer a very practical question: When can we still expect the driver to solve the situation, and when does the car need to help?

The project is led by Chalmers University of Technology, with Volvo Car Corporation as project partner. It is funded by Vinnova, has a budget of SEK 1 million, and runs from August 2026 to May 2027.