BMW calls for better ways to assess driver assist technologies
Speaking about crash avoidance technologies at the ESV conference in Seoul, BMW’s head of safety, Klaus Kompass called for better ways to assess cars’ safety. Kompass also questioned whether continuing to increase requirements for crash protection could be diverting resources from better solutions.
Giving protection for pedestrians and cyclists as an example, Kompass said: “There’s no doubt that the existing requirements for pedestrian protection are absolutely necessary and essential. Around 50% of fatalities are vulnerable road users – pedestrians and cyclists – but the question remains whether pop-up hoods or pedestrian airbags are the right way to address these issues. Collision avoidance could be even more effective.”
BMW favours technologies that make the most of the driver’s intelligence and capability, assisting them to make the right decisions at the right time. “We have a more or less intelligent person sitting behind the wheel,” said Kompass. “We have to consider younger and older drivers, impaired drivers, distracted drivers too – the human factor is important – but we still think the driver is the best sensor we have available.”
The company prefers driver warnings to systems that intervene directly. Autonomous systems that override the driver, Kompass argues, lose a lot of the benefit that could be gained.
It’s a sensible point, but not how Euro NCAP sees it. The safety organisation will start assessing driver assist systems from next year. It will award points to systems that warn drivers and help to reduce the severity of crash, and systems that intervene to prevent a crash completely will score more highly.
BMW’s criticism is that a car with a sure-fire driver warning could never score as many points as one that didn’t warn and just hit the brakes, which could in theory cause more accidents. It’s hard to prove that, however. Testing an autonomous braking system against a dummy car is much more reliable.
Kompass suggests that instead of awarding points for specific features, independent consumer testing should take a different approach. “The current methods for quantifying the benefits of safety features rely on retrospective analysis of accident statistics, looking back in history,” he said. “It will take years to prove whether lane departure warning systems, for example, lead to fewer accidents, but Euro NCAP has selected it and offers incentives for cars equipped with this feature based on, to be a little bit provocative, not much more than good gut feeling.”
So what is Kompass proposing? Instead of testing individual functions with driving tests on proving grounds, he’d like to use a computer model predicting the safety benefits of features and functions. It would be a statistical computer simulation, not a test engineer.
“Let us calculate how many lives could be saved and award points or star ratings based on this,” said Kompass. “If we can collaborate to establish a standard way to do this, carmakers would be able to choose which features and functions they develop.”
This is unlikely to come about soon: Euro NCAP has mapped out how it plans to develop its tests to include driver assistance systems in the next five years.
Kompass raises some interesting questions, however. Once autonomous emergency braking is standard on cars and all the systems are able to prevent head-on collisions with vehicles, pedestrians and cyclists, where do we go next? The remaining accident scenarios will probably be more complex. It could be that a driver assist technology that improves safety for a BMW M3 may not be as effective for a Honda Jazz. If car makers can identify issues and solutions in their own accident research, it could be useful for their vehicle’s safety rating to recognise this.
“It will take years to prove whether lane departure warning systems lead to fewer accidents, …Euro NCAP has selected it based on not much more than good gut feeling.”Klaus Kompass, head of BMW safety