Safety first.
Conditional autonomous driving capabilities (Level 3) have become a highly contested area among various players this year. In addition to the new forces that have always boasted cutting-edge technology, luxury car manufacturers have also emerged as a prominent faction.
On July 27th, BMW announced the launch of its Future Mobility Development Center in the Czech Republic for the research and testing of driver assistance systems and autonomous driving technology. They plan to launch Level 3 autonomous driving capabilities by the end of 2023 or early 2024. The corresponding research and development work has also been initiated in China, and they have stated that they will be fully prepared for the implementation of Level 3 autonomous driving in China in accordance with local regulations.
This is much faster than BMW's original plan for the second half of 2025. It is understood that BMW's Level 3 autonomous driving capabilities will be first introduced in the flagship 7 Series models. Within the next ten years, they will also introduce more affordable solutions in other vehicle models.
Prior to BMW, Mercedes-Benz has already announced its intention to introduce the Level 3 autonomous driving system, Drive Pilot, in China and has already begun road testing. Audi, on the other hand, released the Audi A8 model capable of achieving Level 3 autonomous driving as early as 2017, and after collaborating with SAIC Motor, they are expected to make further progress in Level 3 autonomous driving.
These luxury car players, who have always been cautious when it comes to safety-related technologies, are eagerly providing timelines for the implementation of Level 3 autonomous driving. This indicates that they believe Level 3 autonomous driving is relatively mature in terms of technology and has the value for mass production and commercialization. They are increasing their bets on the second half of the intelligent era, making technological advancements in the field of autonomous driving to attract Chinese consumers.
Compared to Level 2 assisted driving, Level 3 represents a significant milestone in autonomous driving. Starting from Level 3, the system takes over the vehicle's driving, allowing the driver to do other things with their hands off the steering wheel.
However, crossing the gap from Level 2 to Level 3 has always been a challenge for car manufacturers and autonomous driving companies.
Although Level 3 technology is relatively mature, many car manufacturers and autonomous driving companies bypass Level 3 and focus on Level 2 or Level 4. The core issue lies in the fact that below Level 3, the driver is in control of the vehicle, while above Level 3, the system takes charge. In the event of an accident, the responsibilities and liabilities are relatively clear for Levels 0-2 and Levels 4-5. However, Level 3, being in a transitional stage, lacks clear responsibility attribution.
In the past few years, car manufacturers such as Xiaopeng, NIO, and Tesla have experienced accidents related to their autonomous driving features, some of which have even resulted in passenger fatalities. However, these car manufacturers have been difficult to hold accountable, as under existing regulations, the driver is primarily responsible for accidents.
Even Huawei, when promoting its HUAWEI ADS 2.0, only claims that it is approaching Level 3 in terms of advanced intelligent driving.
Players in the industry dare not cross this gap. In the absence of clear laws and regulations and a lack of clarity in liability allocation, any accident would be a devastating blow to the entire automaker or autonomous driving company.
With the support of relevant departments, Level 3 autonomous driving is expected to overcome this challenge that has plagued the industry. In late June, Vice Minister of Industry and Information Technology, Xin Guobin, stated that they support the commercial application of conditional autonomous driving (Level 3) and higher-level autonomous driving functions, and will release a new version of the intelligent connected vehicle standard system guidelines. Currently, L3 autonomous driving policies have been gradually implemented in Germany, the UK, Japan, and South Korea. In terms of accident determination, it is clear that car manufacturers and insurance companies bear the main responsibility unless the driver engages in illegal behavior.
It is expected within the industry that after the policies and regulations are gradually implemented, L3 and even L4 autonomous driving will enter a stage of small-scale mass production in the next three years.
Sinodata Securities predicts that intelligent vehicles will follow Moore's Law, and the long-term cost of intelligent driving hardware will decrease significantly. The BOM cost of L3 intelligent driving perception hardware, including LiDAR, is close to 11,000 yuan and is expected to be around 7,500 yuan by 2025, with a cost reduction potential of over 30%. This will drive down costs.
Autonomous driving will then enter a positive cycle. Further cost reductions will benefit the scale of car manufacturers, allowing them to obtain massive amounts of data and continuously collect and train data to accelerate the iteration of intelligent driving solutions, thereby improving the driving and passenger experience. This will drive more consumers to choose models with stronger autonomous driving capabilities.
According to insiders from leading autonomous driving companies, the hardware in many car models is already capable of supporting vehicles to achieve L3-level or higher autonomous driving. With future regulations gradually being relaxed, vehicles will be able to upgrade from the current advanced driver assistance systems to L3-level autonomous driving, providing a more complete autonomous driving experience. The next three years will be a stage of intelligence development from 1 to 10, reshaping the current market landscape.
Li Zhenyu, Senior Vice President of Baidu Group and President of the Intelligent Driving Business Group, pointed out that by 2026, cars without autonomous driving capabilities will find it difficult to make it onto users' car shopping lists.
However, whether intelligent vehicles can smoothly transition to L3 level and be successfully implemented cannot bypass safety. In addition to the safety of software and hardware, various factors such as environmental changes and disturbances, as well as driver misuse of vehicle functions, will all affect the "safety" of autonomous driving. In terms of safety, there is still much work to be done by industry players.
For example, Mercedes-Benz's Drive Pilot system is currently only available for S-Class models, and BMW's autonomous driving solution is only available for the 7 Series, indicating that widespread application is still a long way off. Even L4-level Automated Valet Parking (AVP) is progressing faster than L3.
Luxury car manufacturers are rushing to occupy the high ground of autonomous driving to avoid falling behind in the second half of the intelligent era. Whether this path can be successfully navigated will become the key to their success in the second half. This still requires extensive safety verification.