What happens with advanced driver assistance systems in the event of a fault – or an accident? New insights from a study by TÜV Rheinland and the UK’s TRL.

Technical expertise and road tests

Lane-keeping and reversing assistance systems, and automatic emergency braking: from 2024, EU law will require new vehicles to include a whole series of advanced driver assistance systems. Sounds good at first. Advanced driver assistance systems do save lives, after all. And the more systems we have that help us as drivers, the safer we are when we’re getting from A to B. At least that’s the case if they work properly and we use them as intended. At TÜV Rheinland, we asked ourselves a simple question: what happens to these systems over a long period of time? Do their sensors still work ‘as new’ after many years on the road? And how do substandard repairs or accidents impact their functionality?

For this project, we took lane-keeping assist systems as our test case and joined forces with the UK’s Transport Research Laboratory (TRL) to look for answers to these questions. We also compiled a knowledge base from the latest published research and spoke to other experts in the field as well as organizations working in the industry. Most importantly of all, we also conducted trials on a test track to experience the practical effects of various scenarios on lane-keeping assist systems at first hand.

More safety – but more research is needed

So what did our study find out? In a nutshell: we remain convinced that advanced driver assistance systems make an important contribution to improving safety on our roads. Yet there are also several reasons for suggesting that these systems can actually increase risks to traffic in certain situations. As of now, we still don’t know enough to be sure and more research is needed – ideally combined with regular testing of these systems.

Let’s look at this in more detail. What DO we know, for example? Well, the sensors that are located behind the windscreen of a vehicle are very important for lane-keeping assist systems in order to work as they should. If the vehicle windscreen has to be replaced, then the sensors have to be recalibrated. Every time. Otherwise, the lane assist system could well fail to identify the lanes correctly as it should.

Exhibit one: incorrect calibration after replacing the windscreen

The problem is, sensors are already often being calibrated wrongly – or not at all – after windscreen replacements. From our research, we can make the following estimates. In 2019, about 2.9 million windscreens that were fitted with camera sensors were replaced in the EU. By 2029, this number is very likely to be around 18 million. Based on this, we estimate that between 1.8 and 7.8 million of the sensors fitted behind windscreens could be wrongly calibrated – if they are calibrated at all (some 10 to 44 percent). Depending on the specific scenario, this calibration issue could prevent the lane assist system from working properly in up to 1 in 5 cases.

Infographic LKA Windscreen

Of course, ‘could’ is not the same as ‘will’. So, to get some more reliable data, our Future Mobility Solutions Team took a hands-on look at how lane-keeping assist handles certain types of scenario. Our trials used a modified test vehicle equipped with a state of the art lane-keeping assist system. And then we really got to work. We simulated damage to the windscreen near to where the cameras were placed, for example, as well as incorrect camera calibration following a windscreen replacement job. Last but not least, we also interfered with in-vehicle data while the car was in motion.

Simulated stone chipping and deliberate data dropouts

Having artificially aged and modified our test-vehicle, we then went to the ZalaZONE Automotive Proving Ground test track in Hungary. During our test drives, we compared the performance of the test vehicle as it navigated different sections of the track (bends and straights). We also focused on situations where no indicator lights or other warning systems had been activated. With the simulated stone chipping affecting the windscreen, for example, we found that the lane assist system would perform less reliably and would even switch off without warning on rare occasions. The assistance system also failed to warn us or even react when we decided to drive over the lane markings. When we deliberately disrupted the vehicle’s on-board communication network during the drive, the system shut down immediately, as it is intendet to.This resulted in the steering wheel suddenly and forcibly moving towards the centre position – a rather unwelcome surprise for a regular driver.

So that was a brief look at the insights from our road trials. And in my opinion, we now need a lot more of these kind of insights. This is especially true because the use of these systems is increasing. While this is to be welcomed in principle, we should not leave the human factor out of the equation. As drivers, we know that these systems are literally intended only to assist us and do not absolve us of our responsibilities. As advanced driver assistance systems become more common, however, I fear we will be increasingly encouraged to let our responsibilities slide. And dangerous situations can then arise if the system then stops working as we expect it to. As an example, the unexpected failure of a lane assist system becomes critical at the moment when the driver is not fully concentrated on the road in front of them or has not got his hands tight on the steering wheel  because they assume the lane assist will handle things.

“The long-term capabilities of any technical system can only be demonstrated by subjecting it to routine maintenance and technical inspections.”

Assistance systems are now much more common

To obtain more data in this field, further studies are now required. One possible focus here would be the ancillary assistance systems like emergency braking systems or other kinds of systems found in the small car segment. I believe this is needed with some urgency because technical developments and new regulatory frameworks are both working to accelerate the rollout of advanced driver assistance systems. As said at the outset, this is very good in terms of improving road safety, but the long-term capabilities of any technical system can only be demonstrated by subjecting it to routine maintenance and technical inspections. Accordingly, independent third parties like TÜV Rheinland will also need access to system data and other details as part of routine vehicle roadworthiness testing.


Rico Barth

Rico Barth

Head of Competence Center Connected and Automated Driving

Rico Barth leads the Competence Center for Connected and Automated Driving within TÜV Rheinland. The graduated engineer with many years of professional experience in the automative sector develops new services as well as validation and testing processes for ADAS and Autonomous Functionalities. Furthermore he is active in various test fields for software and diagnostics, especially in the area of sensor technology, vehicle in the loop approach and connectivity challenges.

More Posts


Field testing a hydrogen car for a day

So how does a hydrogen car actually drive? What should I bear in mind when refueling? Report on a hydrogen car gained during a field test.
Industrie Roboter

Industrial Applications and Safety

Robots – it is impossible to imagine production and logistics without them. Various norms and test standards ensure greater safety when using them.

Ransomware – Kidnapping in the digital age

More and more companies are becoming victims of data kidnapping. A study reveals the need for advice. How to protect against ransomware attacks.



Submit a Comment

Your email address will not be published. Required fields are marked *


No one likes popups. But you’ll like our newsletter.

Get remarkable articles on digitization, modern life, energy and technology.