Conflict between Level 3 Self-Driving Technology and Safety Testing Organizations

By Jeff Lee 2022-09-15

ADAS (advanced driver assistance systems) or L2 driver assistance systems have been prevalent in the global auto market for several years. Using L3 to surpass rivals seems to be inevitable R&D efforts for automakers to make in commercial competition. For one thing, it is the myth about numbers (3 is bigger than 2), and for another, it tests the regulatory stewardship of government regulatory agencies in various countries in terms of the introduction of new technologies. I have always had reservations about the basis the theory of operation for the L3 system is established on, not only because L3 autonomy is nothing more than a transitional product, but also because L3 autonomy constantly tests the driver's emergency response capabilities, which is a condition we do not need to experience when driving today.

    Since the United Nations passed UN Regulation No. 157 (UN-R157) on L3 automated lane keeping systems (ALKS) early last year, international automakers such as Mercedes-Benz, Toyota and Audi have launched related products one after another. The US, the UK, and Japan have also promulgated relevant regulations to encourage L3 vehicles to hit the road. However, the Insurance Institute for Highway Safety (IIHS), known for its stringent vehicle safety tests, has recently proposed new safety rating standards for vehicles with partial autonomy (also known as L3). To obtain a good score, the partial-autonomy technology must keep the driver’s eyes on the road at all times and their hands either on the steering wheel or always ready to grab it. If the driver fails to meet these requirements, the car must send escalating alerts or activate emergency procedures. The reason for setting the strict standards is simple... People become less alert as self-driving functions improve. They are a lot more alert when driving by themselves. The Mercedes EQS, which went on sale in Germany this year, limits the conditions where the L3 system takes control of the car, but the L3 system allows the driver to keep both hands off the steering wheel and do other things in an ODD (operational design domain) environment under the right conditions (and these are the conditions permitted by German law for L3 vehicles). In the event that an unmanageable situation develops, the car will alert the driver to resume driving. If the driver does not respond within 10 seconds, the L3 system will automatically bring the car to a controlled stop and turn on the hazard warning lamps. What is troubling the researchers is that it takes about 10 seconds on average for human beings to shift concentration from something else back to the road after receiving the alert and grabbing the steering wheel. Even though the speed of L3 ALKS vehicles is currently limited to 60 km/h, the car will have moved forward more than 100 meters in 10 seconds, and any incident may occur in those 10 seconds. So, a "controlled stop" alone does not suffice. What L3 needs is a "safe stop", that is, an "emergency maneuver module", but researchers worldwide have not proposed a solution yet.

From the above, the controversy about L3 ALKS technology lies in whether the driver can take over the "stable" control of the vehicle in an emergency. Can all drivers (including the elderly and beginners) manage to refocus and resume driving in time after taking their hands off the steering wheel, as permitted by German law? Or should drivers keep their hands on the steering wheel as specified by the IIHS new standards so as not to react too late? The US and Germany apparently have different perceptions of L3 safety requirements…

  • Germany aims to encourage technological development so that automakers can use new technologies as a marketing tool to appeal to consumers who are willing to be the first to try out new technologies.
  • The US aims to ensure road traffic safety and does not allow for any ambiguity in the transfer of control of a vehicle in motion.

If the introduction of a new technology entails risking personal safety (more risks than in the current situation), then I think the feasibility of commercializing the new technology must be evaluated from several perspectives…

  • Does L3 unequivocally improve driving safety and effectively reduce driver fatigue throughout the driving process compared to L2 (current status)?
  • How useful is L3 when speed is limited to 60 km/h on highways or expressways? Does it pose a greater potential risk to other road users who drive much faster?
  • How often does L3 require the driver to take over when the vehicle enters a congested urban road? Every time the driver takes over, it means a new risk to road traffic safety for sure.
  • The L2 system, with the human driver as the main driver assisted by the carputer, must be equipped with a DMS (driver monitoring system) and a data logger to monitor and warn the human driver before danger occurs and to determine the responsibility after an accident. I don't think it has to take the upgrade to L3 to add the two devices; otherwise, consumers are bound to mistake the addition of these functions for improved driving safety.     

When assessing the commercialization of L3 and road traffic safety for all road users from the foregoing perspectives, I tend to support the IIHS standards. The introduction of a new technology that is transitional in nature (L3 is not the ultimate solution for the future of self-driving) must ensure that consumers and the community enjoy the new technology and recognize the original purpose of its development (to make driving safer). Pulling up the seedling kills it, not helping it grow. If a technology is rejected by the public because it is launched prematurely, such a hasty decision will hamper the timetable for the implementation of L4/L5 autonomous driving.