Air Liquide’s technology chosen for Daxing hydrogen station in Beijing, the world’s largest by refueling capacity
Toyocolor to supply SK Innovation with Lioaccum conductive CNT dispersions for Li-ion cathodes

Study finds advanced driver assistance systems not always reliable in long-term operation

There could be an average of around 790,000 risk events per year on roads in the European Union (EU) in 2029 that are attributable solely to reduced performance of lane-keeping assistant systems, according to a study on the “Lifetime performance of advanced driver assistance systems” by researchers from TÜV Rheinland and the British Transport Research Laboratory (TRL). Risk events are failures in the system that reduce performance.

For the study, the experts used lane keeping assistants (LKA) as an example to investigate the specific impact of age-related wear and tear, damage to the system, and accidents or a lack of calibration of cameras when replacing windshields on the function of assistance systems.

The study comprises two parts. First, the authors compiled current findings based on publications and exchanges with other experts and industry organizations. Second, the experts conducted driving tests on a test track to practically investigate various scenarios.

Advanced driver assistance systems protect lives. From our point of view, it makes sense and is absolutely appropriate that a number of assistance systems will be mandatory equipment for new vehicles in the EU from next year, including lane keeping assistants, emergency braking functions or reversing systems.

Assistance systems must function reliably for many years. With our study, we have gained initial insights into the circumstances under which lane-keeping systems might function to a limited extent only—and into the consequences this may have for road safety.

—Dr. Matthias Schubert, Executive Vice President Mobility at TÜV Rheinland

Schubert believes that too little is known about how accidents, improper repairs or wear and tear affect the functionality of assistance systems and thus road safety in the long term.He therefore advocates conducting further studies on the long-term reliability of the assistance systems and their wear and tear.

Depending on the scenario, the current study by TÜV Rheinland and TRL showed that the estimated number of annual risk events due to malfunctions of the systems could even be as high as 2.3 million on average.

As part of the study, TÜV Rheinland experts took a modified test vehicle equipped with a state-of-the-art lane keeping assistance system on the Zalazone test track in Hungary. For example, TÜV Rheinland experts simulated damages on the windshield in the area of the LKA camera, incorrect calibration of the cameras after replacing the windshield and interruptions in data communication in the vehicle while driving. Furthermore, components were artificially aged. In one scenario, they also made changes to the chassis.

During the test drives, the experts compared how the modified car behaved in specific road sections (curved and straight). The focus was on situations in which neither indicator lights nor other warning systems were activated.

The experts observed, for example, that the function of the LKA deteriorated in simulated stone impacts in the windshield and that in rare cases the LKA switched off without warning. The experts were also able to detect driving over the lane markings without warning or reaction from the system. When contacts in the car’s data line were intentionally interrupted while driving, the system deactivated immediately; the subsequent abrupt return movement of the steering wheel toward the center position can take the driver by surprise.

A risk event can occur, for example, when an aged lane keeping assistance system switches off as intended because it can no longer “see” properly in certain situations due to damage to the windshield.

The increasing prevalence of advanced driver assistance systems means that we rely on them more and more. This happens unconsciously—even if the systems are actually only supposed to relieve us and the responsibility always remains with us as the driver.

The spontaneous deactivation of a system becomes problematic if the driver is not fully concentrated at that moment or does not have his or her hands firmly on the steering wheel at all because he or she has relied completely on the system. To put it differently, there are situations that drivers experience as a malfunction, even though the assistance system is working properly.

—Rico Barth, global head of the connected and automated driving competence area at TÜV Rheinland

Technical development and changes in legal regulations will quickly lead to the widespread use of advanced driver assistance systems. For example, the study’s authors estimate that some 9.7 million windshields equipped with cameras are likely to be replaced in the EU as early as 2025, up from just two million in 2019.


The quality of these millions of repairs also has a significant impact on the proper functioning of assistance systems, components of which are installed in the windshield.

In our view, our study has confirmed that only regular maintenance and technical inspections will demonstrate how well a technical system functions in the long term. Among other things, access to system data for independent third parties such as TÜV Rheinland as part of the periodical technical inspections is crucial to achieving this. The correct functioning of an assistance system when new can be seriously affected by even minor accidents or faulty repairs.

—Matthias Schubert

According to the German Federal Motor Transport Authority, the average age of registered passenger cars in Germany is currently 9.8 years and rising steadily. Schubert therefore believes that further findings should now urgently be obtained on how to ensure the reliable functioning of advanced driver assistance systems throughout their entire lifetime. For example, a further study could look at additional assistance systems such as predictive emergency braking systems or assistance systems from the small car segment.



Of course the answer is better testing and maintenance, but this provides even more significant headwinds for FSD than for 'normal' cars, as there is no driver to blame inadequate sensor performance on when it goes awry.

That might make insuring an older FSD car problematic.


Interesting how 'self driving' has morphed into what it really is...'driver assist.' Me thinks FSD will be a long time coming, if ever!


They'll have to go back to a depot for recalibration etc.
Or they'll have to design them so that they can be operated at non orthogonal angles after calibration.
Humans can operate behind very messy windscreens - it kills me to sit beside someone who won't give the wipers a buzz to clean the windscreen, but they can drive OK.
Machines - less so.
Bigger problems would be that they can't reattach the camera at all, rather than they put it on crooked.
@Lad, the probably will get FSD, certainly in somewhat controlled environments - Autobahn's - yes, LA, yes, Naples or Delhi, no. (This is just a guess - I've never been to either place).

The comments to this entry are closed.