Is it stable? L3 is about to be put into large-scale use, and LiDAR companies are the first winners-EEWORLD

Collect

In China, we are one step closer to the large-scale implementation of L3 autonomous driving.

With the update of domestic pilot policies, more and more car companies have obtained autonomous driving test licenses in some cities, and the number of L3 autonomous driving vehicles that can be tested on the road is increasing rapidly.

One of the important nodes is that not long ago, the Ministry of Industry and Information Technology and other four departments jointly issued the "Notice on the Pilot Work of Intelligent Connected Vehicle Access and Road Passage", announcing that 9 companies have become the first "Intelligent Connected Vehicle Access and Road Passage Pilot Consortium".

The nine shortlisted companies are FAW, SAIC, GAC, Changan, BAIC BluePark, BYD, NIO, SAIC Hongyan and Yutong Bus.

In addition to the nine car company consortiums announced this time, since last year, many domestic and international car companies have obtained L3 test licenses in Beijing, Shanghai, Chongqing and other places. Brands that are actively investing in the field of intelligent driving, such as BMW, Mercedes-Benz, Zhiji, Changan, Arcfox, Deep Blue, Avita, Seres, and Zeekr, are all on the list.

The timetable for the implementation of L3 autonomous driving in China is becoming increasingly clear, and automakers have also begun to propose clear launch dates for L3 autonomous driving models.

Li Xiang, founder and CEO of Ideal Auto, announced at the just-concluded China Automotive Chongqing Forum that Ideal will mass-deliver supervised L3 autonomous driving to users as early as the end of this year and at the latest in the first half of next year.

Various sources of information also show that international luxury brands such as Audi, Mercedes-Benz, and BMW are also accelerating the mass production of high-end intelligent driving. Their L3 autonomous driving research and development projects, which have been brewing overseas for many years, may even be first implemented on a large scale in China.

BMW is the first company in China to obtain a Level 3 autonomous driving test license. Its Level 3 autonomous driving technology suite will integrate a new lidar system. The domestically produced new BMW 5 Series is technically capable of expanding to Level 3 autonomous driving.

Mercedes-Benz, the first company in the world to launch an L3 autonomous driving model, has made it clear that it is promoting the L3 system Drive Pilot to enter the Chinese market after the United States and Germany. According to 36Kr Auto, the CLA model jointly developed by Mercedes-Benz and Momenta will adopt a combination of LiDAR + millimeter-wave radar + camera.

Mercedes-Benz self-driving test car photographed by overseas netizens, equipped with laser radar

According to Counterpoint Research, by 2026, the number of L3 autonomous driving cars on the road in China will reach 1 million, and their shipment share will account for 10% of the market in 2028.

L3 autonomous driving means that the main responsibility for driving shifts from the driver to the car manufacturer. At the same time, this will bring about a change in the business model for the entire automotive industry. Zhijia's potential subscription payment model may change the business model of car manufacturers and the Zhijia supply chain.

L3 will be a key new milestone in the development of intelligent driving.

So, with the coming of L3, what changes will it bring to the automotive industry? In this process, looking at the entire intelligent driving industry chain, who will be the first winners?

L3 reconstructs the industrial chain.

Even change the intelligent driving business model

Based on the standard of "whether the driver needs to be in the driving state after the autonomous driving function is turned on", autonomous driving is divided into assisted driving and autonomous driving with L3 as the dividing line. The change in driving responsibility is the core of L3. Strictly speaking, only intelligent driving at L3 and above can be called "true autonomous driving".

As a key node in the development of intelligent driving, L3 will promote the popularization of the concept of intelligent driving to a greater extent. The shift in driving responsibilities and user attention will also greatly enhance the practicality and usage scenarios of intelligent driving.

In a highly inward-looking market, getting consumers to pay for intelligent driving systems is one of the major challenges currently facing automakers.

In addition to the brand influence of car companies, L3 may also change the business model of smart driving and strengthen users' willingness to pay for smart driving. For example, Tesla's FSD, priced at 64,000 yuan, can account for about 20% of the overall price of its models, and stronger and more advanced functions will likely increase users' willingness to pay.

In sync with domestic car companies, Tesla is also accelerating the development of FSD with super-high investment, and plans to invest $10 billion to accelerate the development of FSD in 2024. As FSD is about to enter China, the latest L3 systems of domestic manufacturers will likely compete with FSD.

As L3 autonomous driving technology matures, intelligent driving technology will gradually be popularized in more car models and markets, providing a broader space for the promotion of software payment models.

L3 level puts forward new requirements for LiDAR

However, before that, how smart driving can cross the gap from L2 to L3 is the biggest challenge facing the automotive industry. After all, once L3 has safety problems, it will not only harm the driver himself, but the car company will also need to bear more responsibility.

This also requires the autonomous driving system to have a more thorough redundant safety design, including redundant chassis, perception, computing, mapping, V2X, etc., and the vehicle's software and hardware systems also need further iterations to be ready for L3 driving.

L3 policy documents mention sensor requirements

Currently, there are two main routes for the development of intelligent driving perception systems in the industry. One is the pure visual technology route, and the other is the multi-sensor fusion technology route, mainly based on lidar + camera + millimeter wave radar.

Most domestic manufacturers have adopted solutions with lidar.

Taking the Huawei-supported Wenjie product line as an example, Wenjie's M5 and M7 models, which were modified this year, have significantly increased the configuration proportion of intelligent driving systems with laser radar.

In the past two years, with the sales of smart driving models, AEB has attracted strong attention among car owners.

LiDAR can detect directly by actively emitting lasers and does not rely on ambient light. Therefore, using LiDAR data for AEB function development can greatly enhance AEB safety at night. An international paper pointed out that at night, the average perception accuracy of LiDAR data is three times that of pure vision.

AEB improves driving safety and can save a large number of lives every year. In the EU, this number is about 1,000 per year. According to a media survey, 92% of consumers will give priority to vehicles with AEB when buying a car.

In the Chinese market, with the hot sales of NIO's entire series, Ideal L series, Hongmeng Intelligent Driving World, Zhijie and other series of models, users' awareness and enthusiasm for LiDAR have also greatly increased. Many consumers are willing to pay for LiDAR for higher-level intelligent driving functions and safety.

The demand for large-scale implementation of L3 has accelerated and strengthened the majority of automakers’ choice of intelligent driving systems with LiDAR and multi-sensor fusion. For example, among the six automakers that have obtained test licenses, five have installed LiDAR in their intelligent driving system solutions.

In general, international luxury brands represented by BBA and most domestic automakers have reached a consensus that higher-level intelligent driving, including L3, will strengthen the role of LiDAR. With the implementation of L3 policies, the speed of LiDAR installation will inevitably be accelerated.

From the rise of L4 in 2016, to the mass production of L2+ in 2020, to the implementation of L3 policy this year, automakers' demand for lidar has continued to evolve.

In the first stage, L4 autonomous driving requires high performance, but not necessarily automotive-grade LiDAR products;
In the second stage, with the fierce competition in the mass production of L2+ intelligent driving and smart electric vehicles, OEMs have a strong need for products that can be mass-produced, automotive-grade, and cost-controlled.
In the third stage, the demand for lidar has increased significantly again. The overall system requirements of L3 are expected to be 100-1000 times safer than human driving without human involvement.

Xue Fuming, member of the board of directors and chief technology officer of Mercedes-Benz Group AG, who is responsible for R&D and procurement, has previously talked about the requirements of intelligent driving systems for lidar:

If we only consider urban roads or urban ring roads, and the driving speed is about 50-70 km/h, then the lower-cost LiDAR can meet the requirements;
If you want to drive a vehicle at 130 km/h, the requirements for LiDAR will change (increase significantly).

Taking the high-speed scenario as an example, one of the important threats for the L3 system to achieve such a high level of safety is small-sized targets at a long distance, such as fallen cargo, sandbags or motorcycle helmets on the highway.

High-speed spillage scenario, a typical small target

This means that L3 autonomous driving requires LiDAR to have higher angular resolution and longer detection distance. The LiDARs currently used in L2 intelligent driving systems may not be able to meet such high requirements, and OEMs may have to reselect higher-performance LiDAR products for L3.

At the same time, some leading intelligent driving system solution companies are stepping up their recruitment of lidar-related talents, involving the design and development of perception algorithms, as well as the application and optimization of algorithms in different levels of autonomous driving products.

L3 is coming, who will be the biggest winner?

Just as the penetration rate of the new energy industry has increased to more than 50% this year, lidar is also accelerating its crossing of the gap.

According to statistics from Gaogong Intelligent Automobile Research Institute: In the new energy market with a capacity of more than 150,000 vehicles, the installation volume of LiDAR in vehicles will reach 10.4% in 2023.

It is estimated that there will be about 6 million new energy vehicles with a capacity of more than 150,000 units in 2024; the total shipments of LiDAR companies in the entire market will be about 600,000 units in 2023, and the conservative forecast of each company is at least 1 million units in 2024, which will make the penetration rate of LiDAR exceed 16%, and it is about to cross the gap and penetrate into the mainstream population.

The installation of LiDAR has penetrated into the mainstream population. Even if the penetration rate is only 16%, it has already crossed the gap for the LiDAR industry.

But in fact, the penetration rate of lidar is faster than most people think.

The latest data shows that in April this year, the penetration rate of laser radar in the new energy market with a sales volume of more than 150,000 yuan was 18.4%; from January to April, the penetration rates of highway NOA and urban NOA in new energy passenger cars reached 20.8% and 12.2% respectively.

For lidar companies, this is undoubtedly an opportunity to enter the fast lane of industry development and accelerate their market capture.

At present, the main players in the domestic laser radar companies include RoboSense, Hesai, Huawei, Tudatong, and Yijing Technology.

Among them, RoboSense's delivery volume continued to rise thanks to its outstanding performance in Xpeng, Hongmeng Intelligent Driving and other car series.

In the first quarter of this year, its total sales of lidar products reached 120,400 units, of which 116,200 units were lidar products in the ADAS field, a year-on-year surge of 542.0%.

During this year's CES, Sagitar also launched a new generation of lidar M3. The M3 adopts a 940nm laser transceiver solution and can achieve a ranging capability of 300 meters @ 10%. This is Sagitar's latest product for L3 autonomous driving applications.

Hesai, another leading lidar company with strong shipments, has added at least nine new automakers as its partners in the past year, including the popular Xiaomi, one of the top three domestic automakers Great Wall Motors, new powerhouse Leapmotor, as well as the country's largest new energy OEM and mainstream European and American joint venture automakers.

As of now, Hesai has obtained orders for nearly 70 vehicle models from 18 OEM customers in the original equipment mass production market.

In April this year, Hesai released the new generation of LiDAR products ATX and AT512, both equipped with Hesai's fourth-generation chip architecture. AT512 has a performance of more than 10 million points per second, which is 8 times the performance. It is a flagship product that helps OEMs mass produce L3 autonomous driving.

Huawei also released its updated lidar product D3P during the Beijing Auto Show in April this year, and will mass-produce the next-generation product D5 with higher specifications and better cloud computing in 2025. The main optimization direction of D5 is also to "greatly improve the detection capability of small targets" to support the mass production of L3 autonomous driving.

The latest trends of the leading LiDAR companies indicate that with the opening of the L3 stage, LiDAR is returning to the performance competition track. Longer-range and denser LiDAR products that serve stronger autonomous driving systems have become the target of competition among the leading companies.

The industry is not just a competition of cost reduction and scale expansion. L3 will bring intelligent driving systems, including lidar, back to the forefront of technological innovation and breakthroughs.