▲ Professor Heo Geon-su of Hanyang University’s Department of Future Automotive Engineering is giving a presentation.

Cruise autonomous driving 145,000 km, Korean drivers 1 accident per 322,000 km
It is time to develop core technologies required for Level 2+ technology advancement and Level 3 and above.

Experts have argued that since not all core technologies for autonomous driving level 3 or higher have reached the commercialization level, continuous development and technological internalization of technologies for level 2+ technological advancement and level 3 or higher autonomous driving are necessary.

Professor Heo Geon-soo of Hanyang University's Department of Automotive Engineering appeared as a keynote speaker at the KADF (Korea Autonomous Developer Forum) 2023, co-hosted by Ansys and Morai, on the 26th to present trends in autonomous driving technology and the progress of commercialization stages, and said that autonomous driving at Level 3 or higher has not yet been developed to the level where it can be driven.

While accident frequency has been considered a criterion for securing the safety of autonomous driving technology until recently, Cruise reported that autonomous driving was disengaged about once every 145,000 km, and an accident usually occurs once every 65,000 km.

In the US, when people drive themselves, one traffic accident occurs every 800,000 km, and in Korea, one traffic accident occurs every 320,000 km.

The result is that autonomous driving has not yet demonstrated stability that surpasses that of humans.

Contrary to expectations that autonomous driving would eliminate accidents, data released by the U.S. Department of Transportation shows that even if autonomous driving becomes perfect, traffic accidents caused by humans would actually only be reduced by about one-third.

Additionally, Toyota of Japan analyzed that even if Level 2 is fully implemented in the United States, traffic accidents will only decrease by about 50%.

Currently, the autonomous driving industry has not been able to give a clear answer to the question of whether autonomous driving can guarantee that it will not cause traffic accidents in a situation where humans do not cause traffic accidents. Tesla is continuously analyzing the judgments and patterns of human drivers and autonomous driving to solve this problem.

Regarding the direction of the autonomous driving business, Professor Heo said that it is expected that the appropriate business model will be to focus on advancing Level 2 as Level 2 technologies are introduced one by one, and that it is a time to continuously develop the core technologies required for Level 2+ technology advancement or Level 3 or higher.

GM is enhancing the capabilities of Level 2, called Super Cruise, to Level 2+, called Ultra Cruise, and the difference between the two is that Level 2+ utilizes lidar and map data, unlike Level 2.

Toyota is also challenging Level 2+ and is currently trying to implement Level 2+ using a precision map and a sensor unit that covers 360 degrees.

Professor Huh took a critical view of Level 3.

Mercedes-Benz's Level 3 autonomous driving is limited to 60 km/h and is affected by weather conditions such as rain and snow, and has difficulty operating normally in tunnels.
> In the end, there is little difference between the specifications of major OEMs that are promoting Level 2+ and the specifications of Level 3 that Mercedes-Benz is talking about.

Professor Heo claims that there are still many technological hurdles to overcome for levels 3, 4, and 5, where autonomous driving functions must not be lost at any moment and all technologies on the right side of the above diagram must work in perfect synergy.

Professor Huh Geon-soo said, “It is very difficult to make a self-driving car that is safer than humans, and there are many shortcomings in selling a self-driving car equivalent to Level 3,” and predicted that the situation of Level 2+ advancements will continue for at least the next five years.