The gap between expectations and reality is extremely cruel.
As a variety of sensors for intelligent driving, millimeter-wave radar with a relatively low threshold (technology and cost) could have achieved better results in 2023. However, it backfired.
Gaogong Intelligent Vehicle Research InstituteMonitoring data shows thatIn 2023, passenger cars in the Chinese market (excluding imports and exports) will be equipped with millimeter-wave radar (excluding cabin) 2217 as standard930,000, an increase of only 23 year-on-year54%, compared with the growth rate of the same period last year, down nearly 8 percentage points.
Among them, BYD was originally expected to place orders for 5R on a large scale last year (at that time, it was rumored that the world's major radar chip manufacturers were urgently stocking up, with the order of millions of units), and the actual release was far lower than expected.
According to the data, in 2023, nearly 800,000 new vehicles with 5R configuration will be delivered in the Chinese market, of which BYD's model delivery contribution will only account for 1278%；According to BYD's latest intelligent driving configuration strategy adjustment at the beginning of this year, the follow-up situation will not get better.
According to BYD's public plan, next, models above 200,000 yuan can only provide optional high-end intelligent driving assistance systems, and models above 300,000 yuan will be standard. This means that models below 200,000 yuan, which account for more than 7% of BYD's passenger car sales, will not have a chance with high-end intelligent driving.
The demand for 5R configuration mainly comes from high-end intelligent driving.
In addition, Tesla, Mobileye, etc., continue to strengthen the cost-effective advantages of pure visual perception solutions, coupled with the introduction of LiDAR in urban areas to reduce the configuration of millimeter-wave radar, and other factors, which also bring huge uncertainty to this market segment.
In particular, the cost reduction of car companies has reduced the configuration of the previous 5R to 3R. Taking Huawei as an example, since last year, ADS 2The 0 system has been reduced from the previous 6 millimeter-wave radars to 3 in some models.
Since last year, many leading manufacturers have begun to hope that 4D point cloud millimeter-wave radar can obtain considerable orders in the alternative upgrade cycle. But, obviously, it is not satisfactory either.
So far, in the domestic market, only ZF, Senstech and Foretek have achieved the pre-installation delivery of 4D millimeter-wave radar. From the perspective of car companies, "the current 4D radar is still used as an ordinary radar, which is a new technology, so it is still verifying many high-end capabilities." ”
Similarly, Tesla also confirmed that at present, the company only conducts relevant verification work on the Model S and X models to evaluate the actual performance of the products, and there are no plans to install 4D radar on the Model 3 and Y models.
With the continuous iteration of visual perception technology and the tendency of high-end NOA to lidar, the configuration of millimeter-wave radar, especially 4D radar, is still in the game stage. Industry insiders said frankly.
At present, the industry is still in a vague state about whether 4D millimeter-wave radar will replace the original 3D radar for perception redundancy, or as the main sensor. In particular, the processing of 4D point clouds is a "sweet" annoyance.
Actually, in addition to adding 1d altimetry parameters,The biggest advantage of 4D point clouds is that the cost is much lower than that of LiDAR, but it can also achieve high-resolution obstacle identification and classification. However, the technical barrier has also been raised, "the hardware has become easier, but the software has become more difficult." ”
Previously, ARBE, a listed company on 4D radar in the United States, said bluntly: We are providing a software stack as much as possible to help our end customers (automakers) adapt to this technology as soon as possible. For car companies and Tier1s, it also takes time to optimize software and hardware performance.
In addition, the market is also actively promoting satellite-based architecture sensor configuration schemes, which take full advantage of the computing power redundancy of domain controllers and advanced algorithms for object detection, classification, and tracking, thereby reducing the computing power cost on the device side (previously, 4D radars used expensive FPGAs).
The benefits of this architecture are obvious, and the data of multiple surround 4D radars are processed collaboratively and fused in the first-class domain controller (for example, the intelligent driving domain controller), thereby reducing the data processing requirements on the device side.
At the same time, with the help of the BEV feature map of the radar, it can be fused with the BEV feature of the image to further enhance the confidence of target detection. At the same time, however, the requirements for software development are higher.
As we all know, from the technical upgrade of traditional 3D to 4D imaging, the software algorithm ability has become the shortcoming of millimeter-wave radar companies that relied more on hardware design, development and manufacturing in the past. (Especially from 4D to imaging, the requirements for software are also a qualitative change).
However, architectural innovation will not bring large-scale effects in the short term. In particular, it is more difficult for pure radar first-class vendors to enjoy technical dividends compared with tier1 giants.
How to get car companies to get on the bus first and then buy tickets? Quickly reduce costs and become the only way out for 4D radar companies, especially startups. After all, lidar has played a similar routine in the early days, and many car companies have not completed the development of software before getting into the car.
Public information shows that last year, the delivery of 4D imaging radar** has dropped to 500-1000 yuan, and the lowest has even fallen below 500 yuan, ** is rapidly approaching 3D forward radar. According to a number of business sources, it is expected that the ** will decline by 30%-40% in 2024.
This demand mainly comes from the direct replacement of traditional 3D radar. However, the performance itself will not have much impact on the computing power of the vehicle (for example, increasing the height measurement capability and outputting more point clouds than before).
However, for China's local millimeter-wave radar manufacturers, the time window is shorteningIn particular, traditional giants such as Bosch, Aptiv, and Continental are continuing to make efforts. Local R&D, cost reduction, and product iteration are accelerating.
For example, Aptiv launched the seventh-generation 4D millimeter-wave angle radar last year, which was developed by a local Chinese team and equipped with the first integrated radar chip in China and the industry's first air waveguide antenna solution.
This high-performance corner radar is designed based on a 4-transmitter and 4-receiver domestic radar chip, and adopts the latest anti-jamming algorithm independently designed by Aptiv, which improves the anti-jamming capability by about 30% compared with traditional radar.
In addition, Aptiv is also one of the first companies in the world to propose the concept of sensor satellite architecture. According to the data given by the company, the above architecture can reduce the size of the radar sensor by 70% and the weight by 30%, and at the same time, it is also more conducive to the future OTA upgrade of the sensor.
Bosch has previously launched the fifth-generation 4D millimeter-wave radar Extreme Edition, which can achieve long-distance detection, increase the horizontal viewing angle and further improve the detection accuracy. Compared with the most advanced millimeter-wave radar, the bandwidth is increased by 50%, the point cloud density is increased by 10 times, and the environment modeling is realized by expanding the occupied bandwidth.
Last year, the company began to focus on mass production and delivery of the sixth-generation millimeter-wave radar, just two years after the launch of the previous generation, further highlighting its sense of urgency to compete for the market.
This radar uses the SoC chip and 3D waveguide antenna technology of the latest RFCMOS technology, and is based on AI-based target classification and road edge classification, providing farther and higher resolution recognition performance to better detect low-reflective objects (such as two-wheeled vehicles).
In addition, last year, Continental began to launch its sixth-generation radar product family with 5R solutions in the Chinese market, including one front radar ARS620 and four surround radar SRR630.
From the point of view of parameters, ARS620 adopts 4T4R, a single chip integrates radio frequency and signal processing, the maximum detection distance of 250 meters for motorcycles, has the ability of dense point cloud output, and the short-range angle FOV is 60°.
At the same time, this generation of products adopts a modular design, reducing complexity while greatly improving performance, with a maximum detection distance of up to 280 meters for forward-facing long-range radar and a 360-degree detection range of up to 200 meters for surround radar.
More importantly, compared with the ARS540 (Continental's first 4D radar), the ARS620 is more controllable in cost (the performance is between 3D and 4D), and it is obviously on the path to large-scale operation. High performance and reasonable are the main labels of this product.
According to the plan, the company will also launch the next generation of 4D radar, which also focuses on high performance and low cost, and will be more competitive in the market than the first generation of ARS540.
In Continental's view, in the Chinese market, independent brand car companies, especially the head of the new forces, continue to lead the intelligent, the first business must adjust the product portfolio more frequently than before to keep up with the pace of rapid changes in the industry.
In the next five years, we will continue to increase our investment in China and devote more resources to deepen our local practices, and we are committed to achieving 50% business growth, and will further expand our cooperation with Chinese domestic brands to reach 60% of the relevant business. Aptiv said.
In the view of the Gaogong Intelligent Vehicle Research Institute, entering 2024, the continuous turbulence of the market, the uncertainty of demand, and the fierce competition will test the anti-risk ability of millimeter-wave radar manufacturers, and the pressure is also shifting to China's local manufacturers, especially start-ups.
At the same time, the self-developed millimeter-wave radar in the system of car companies represented by BYD, Great Wall, and Tesla has further squeezed the living space of third-party manufacturers.