After returning again after four years, and superimposed on the May Day holiday, this year's Beijing Auto Show attracted nearly 1,000 domestic and foreign automobile industry companies to participate in the exhibition, with 117 world premieres and 41 concept cars - a large number of new cars have also taken this opportunity to debut one after another.
Compared with four years ago, China's auto market has changed significantly, with the proportion of domestic brand sales surpassing that of overseas brands, and the monthly penetration rate of new energy vehicles exceeding 50% for the first time.
In April alone, the starting price of the new M7 was lowered by 20,000 yuan, NIO launched a replacement subsidy of up to 1 billion yuan for oil vehicles, and FAW-Volkswagen added a cash subsidy of up to 5,000 yuan for a limited time......
The smell of gunpowder in China's auto market continues to spread.
In this context, the middle and high-end intelligent driving functions, which are regarded as the key words to snatch consumers' minds, have also begun to try to explore the waist market at the price of 10-250,000 yuan.
"In the past, an important reason why intelligent assisted driving did not become the top three car purchase decisions of customers was that it was too expensive. ”
He Xiaopeng, chairman and CEO of Xpeng Motors, said at the China Electric Vehicle 100 Conference in March this year, "We see that high-level intelligent assisted driving is about 50,000 yuan more expensive than the general intelligent assisted driving that only provides ACC." If it is only 10,000 yuan or less than 20,000 yuan, we believe that high-level intelligent assisted driving or even unmanned driving will quickly occupy this market. ”
The low-cost route is providing new opportunities for domestic intelligent driving players.
Radar is smaller, more powerful, and less cost--millimeter-wave radar, which was once thought to be replaced by lidar, is also giving new vitality, showing an increasingly important and critical appearance in this "cost-effective" throat war of car companies.
When the popularization of high-end intelligent driving is underway, lidar is no longer the key
High-end intelligent driving functions are facing an embarrassing reality.
McKinsey pointed out in the "2024 McKinsey China Automotive Consumer Insights" report that consumers' interest in various autonomous driving functions is increasing, but the willingness to pay extra for them is declining, especially in first-tier cities.
In order to promote the popularization and application of high-end intelligent driving, the "cost reduction war" of high-end intelligent driving represented by NOA is being fiercely staged.
"The cost reduction and efficiency increase of high-end intelligent driving solutions means that the popularity will be very high. If some high-end intelligent driving functions are only used in high-end models, when it comes to whether to choose them, customers may be more concerned about how high the functions and performance are realized, rather than paying attention to the cost. ”
An industry insider pointed out to the new intelligent driving that the overall trend of the intelligent driving industry is shifting from the original pursuit of large computing power to the competition for cost performance, and it is hoped that the intelligent driving function will be used as the standard configuration on the mainstream models, and the price range of the intelligent driving model will be broadened, and the cost performance will become more and more important.
More car companies have begun to try high-end intelligent driving solutions that "remove lidar". For example, in early April, the newly released S7, the basic version of the intelligent world jointly built by Huawei and Chery, debuted the basic version of Huawei's visual intelligent driving. According to reports, the new scheme does not use lidar, relying on sensors such as millimeter-wave radar and ultrasonic radar, it can realize functions such as pilot assistance, on-ramp and off-ramp, and intelligent parking on highways and urban expressways across the country.
Another example is the latest model of Xpeng Motors, which will be launched in the second half of this year, F57, it is reported that its intelligent driving system will also cancel lidar and instead use visual solutions and millimeter-wave radar.
On the other hand, major suppliers are also launching cost-effective medium and high-end intelligent driving solutions, and the price of 5V5R intelligent driving solutions launched by many suppliers has even been pushed to the level of 3,000 yuan to 5,000 yuan.
At the same time, with the rapid update of autonomous driving algorithms, BEV (Bird Eye View) + Transformer and the use of Occupancy Network (CNN) are gradually replacing the traditional CNN (Convolutional Neural). Network, convolutional neural network), relying on stronger AI computing power and more data samples and visual processing capabilities, the perception ability of the intelligent driving solution can also be greatly improved on the basis of not relying on lidar.
Even if more companies begin to join the team of pure visual routes, they are actually inseparable from the assistance of millimeter-wave radar.
For example, the Jiyue 01 backed by Baidu only needs to be equipped with 11 cameras + 12 ultrasonic radars + 5 millimeter wave radars, and on the basis of BEV+Transformer, the OCC occupancy network technology is updated through OTA upgrades, which improves the recognition ability and scene generalization ability of special-shaped obstacles, and achieves the so-called "pure visual high-end intelligent driving is not inferior to Tesla".
Another example is Chery's new square box model iCAR V23, which was newly unveiled in April, equipped with an 800w front-view camera + 300w surround view camera, 5 millimeter-wave radars, and is claimed to be able to support L2++ high-end intelligent driving, memory parking, high-speed NOA (pilot assisted driving) and other functions.
In addition to the high-end intelligent driving market, most of the L2 intelligent assisted driving functions with lower perception requirements do not need to use lidar when they get into the car. With the popularization and application of ADAS functions in new energy passenger vehicles, the market penetration rate of millimeter-wave radar is still increasing.
According to the data of Gasgoo Automotive Research Institute, in 2023, the standard configuration of L2 level (including L2+ and L2++) in the new energy passenger car market will reach 3.7 million units, with a market penetration rate of more than 51%. Among them, the penetration rate of models in the price range of 20-500,000 has exceeded 85%. In the future, models below 100,000 will also usher in the growth of ADAS.
According to the monitoring data of the Gaogong Intelligent Vehicle Research Institute, in 2022, 17,952,700 front-mounted millimeter-wave radars equipped with ADAS (forward, backward, and blind spots) will be delivered in the Chinese market (excluding import and export), a year-on-year increase of 31.21%, of which the forward-mounted equipment will increase by 25.21% year-on-year, and the blind area will increase by 37.73% year-on-year. By 2025, the total number of millimeter-wave radars will reach 35.32 million, with a compound annual growth rate of 29.90% from 2020 to 2025.
More importantly, from the early single front radar 1R to 1R+2R, 1R+4R, and then to 4D imaging radar, from supporting AEB and ACC, to supporting BSD, lane change assistance, and then to supporting NOA and driving and parking integration, the technology upgrade has also brought more possibilities for millimeter-wave radar on the car.
Technology upgrades, more possibilities for millimeter-wave radar on the car
The original intention of the birth of intelligent driving is to improve driving safety, liberate manpower, and broaden the driving and application scenarios of vehicles, and break through the working conditions that can be operated by people, but due to the limitations of existing perception technology, the current advanced intelligent driving is not enough to achieve all-weather operation under working conditions.
Therefore, in high-level intelligent driving, players have to use multiple sensors to make up for the lack of a single sensor.
At present, the main sensors used in intelligent driving include "visible light camera", "infrared camera", "millimeter-wave radar", "lidar" and "ultrasonic radar", which have their own advantages and disadvantages in practical application.
Specifically, for example, the advantages of millimeter-wave radar are:
(1) High resolution, small size. Since the size of antennas and other microwave components is frequency-dependent, millimeter-wave radar antennas and microwave components are smaller, and narrow beams can be obtained with small antenna sizes.
(2) Compared with optical sensors such as infrared, laser, and camera, millimeter-wave radar sensors have strong ability to penetrate fog, smoke, and dust, and the ranging accuracy is less affected by weather factors and environmental factors, which can ensure the normal operation of vehicles in various daily weather.
(3) One advantage of millimeter-wave radar over infrared systems, which are often used to compare with millimeter-wave radar, is that distance and velocity information can be measured directly.
In the ADAS system, the specific application of millimeter-wave radar can be divided according to the different needs and functions of the vehicle, such as forward radar, backward radar and corner radar according to the different installation positions in the car, and can also be divided into long-range radar, medium-range radar and short-range radar according to the distance of detection.
Such as AEB automatic braking, FCW forward collision warning, LCA lane change assist, ACC adaptive cruise, BSW blind spot monitoring, etc., are all specific applications of millimeter-wave radar in ADAS.
When domestic car companies build a full perception system, they generally rely on semi-solid-state lidar to provide forward long-range measurement functions, and cameras and millimeter-wave radars are generally responsible for the near-range surround perception on both sides of passenger cars.
However, with the development of technology, the application scenarios of millimeter-wave radar in vehicles are constantly expanding, the detection range is getting farther and farther, and the accuracy of measurement is gradually improving.
For example, in 2021, Continental successfully mass-produced the world's first 4D imaging radar ARS540.
In terms of performance, 4D imaging millimeter-wave radar is an upgraded version of 3D millimeter-wave radar, on the other hand, from the cost point of view, the cost of 4D imaging millimeter-wave radar is only 10%-20% of that of lidar.
Compared with the traditional 3D millimeter-wave radar, the vehicle-mounted 4D millimeter-wave radar can not only calculate the distance, speed, and horizontal angle information of the target, but also the pitch angle information of the target, which can then provide information about the environment around the car, and can avoid false alarms caused by manhole covers, road shoulders, and speed bumps.
In addition, thanks to the ability to provide altitude information of the target, capture the spatial coordinates and velocity information of the target around the car, the 4D millimeter-wave radar can also provide more realistic path planning and passable space detection functions.
Another example is the sixth-generation long-range radar launched by Continental in 2021, which has developed the detection range of millimeter-wave radar to about 280 meters, and its surround radar that can detect the surrounding environment of the vehicle body in 360 degrees, with a detection range of up to 200 meters, its performance far exceeds that of the previous short-range radar.
More importantly, as various car companies have landed in urban NOA, BEV perception algorithm model has become the only way to urban NOA, and Continental's sixth-generation radar can not only be used as an intelligent radar to process data in sensors, but also as a satellite radar to process data in the central control unit to make up for the computing power bottleneck of the radar itself.
In layman's terms, BEVs require a large number of radar point clouds, and a large amount of point cloud data cannot be achieved and provided by ordinary radars, while Continental's satellite radars can produce much more point clouds than ordinary radars.
However, the difference between satellite radar and traditional millimeter-wave radar is that satellite radar outputs relatively early data of the radar (relatively early data in the data processing process), so it needs to output these data to the domain controller or related controllers for further back-end data processing through Ethernet, so as to form a large number of point clouds to meet the needs of BEV algorithm for point clouds.
"The large amount of front-end data output of satellite radar provides more effective data for customers' high-computing platform algorithms, thereby improving the KPIs of the entire front-end fusion, target number, target performance and so on. Continental's technical experts introduced the new intelligent driving.
Smaller size, higher performance, lower cost, where is the future of millimeter-wave radar?
At present, from the perspective of the entire vehicle millimeter-wave radar market, it is mainly controlled by some foreign manufacturers such as Germany, the United States, and Japan, among which Continental, Bosch, Denso, Veoneer, and Aptiv (formerly Delphi) are the most famous. In particular, the 77GHz millimeter-wave radar is mainly mastered by companies such as Continental, Bosch, Aptiv (formerly Delphi), Denso, Trina, Fujitsu Ten, and Hitachi.
In recent years, domestic manufacturers have also actively sought breakthroughs in the field of millimeter-wave radar, but due to the late development, domestic manufacturers are still far from foreign manufacturers in terms of stable product development and iteration, vehicle manufacturer adaptation experience, radar performance limit parameter values, and radar stability in special scenarios.
For example, Continental has been deeply engaged in the field of millimeter-wave radar for more than 20 years, and its radar products have been iterated to the sixth generation. Due to the earlier development, Continental's R&D team and products have undergone a lot of testing and optimization, and can still maintain stable radar performance in special scenarios such as corner cases such as tunnels and extreme climatic conditions.
For example, millimeter-wave radar is very sensitive to metal surfaces, which leads to poor performance of products when passing through tunnels, and many domestic manufacturers are quite headaches for this, while Continental can significantly improve the performance of millimeter-wave radar in tunnels by optimizing technical solutions.
However, the self-transformation of millimeter-wave radar has not stopped.
For millimeter-wave radar, the current development in the industry is mainly focused on:1. The integration of chips is getting higher and higher, 2. Waveguide antenna technology, 3. Radar is getting smaller and smaller, 4. High performance of radar and other fields.
In fact, these development trends are inseparable from the core demand of "reducing costs and increasing efficiency" of car manufacturers.
The cost structure of millimeter-wave radar is mainly algorithm + RF front-end + signal processing chip + high-frequency PCB board. Continental told Xinzhijia that by optimizing the antenna structure and adopting the new RF chip + waveguide antenna technology, Continental's product cost is continuously reducing and the performance is also improving.
Continental's sixth-generation millimeter-wave radar sensor
For example, Continental's sixth-generation radar uses Launch on Package (LoP) technology (based on TI chips), which allows electromagnetic waves emitted by the radar to propagate directly from the chip through the air wave direction. This avoids the loss of electromagnetic waves on the PCB, thus improving radar detection performance.
Another example is the optimization of waveguide antenna technology, which can further reduce costs and increase efficiency.
A waveguide is essentially a miniaturized, rectangular tube that can carry high-frequency radio waves. When the inside of a waveguide is filled with a medium, the composition of the filled medium determines the energy loss as the electromagnetic wave propagates through it.
Continental's sixth-generation radar uses an air waveguide antenna, because the medium of the transmission layer and the radiation layer are air, so the loss of the entire transmission and radiation process is very small, and at the same time brings the improvement of radiation efficiency and radiation range, which further brings benefits to the gain (energy intensity) and pattern (energy coverage) of the antenna.
"By reducing the loss of transmitted and received signals, air waveguide antennas can be detected more sensitively and over longer distances. Except for the difficulty of the current technology (only a few manufacturers have mastered this technology), waveguide antennas have better performance than microstrip antennas in all other aspects (cost/performance/volume, etc.). Continental noted.
At the same time, combined with the Launch on Package (LoP) technology of the sixth-generation radar, the electromagnetic waves emitted by the radar can propagate directly from the chip through the air wave guide.
In addition, there are anti-jamming technology used in the sixth generation radar, CCM frequency modulation signal innovation, and super-resolution algorithm technology...... With more than 20 years of experience in R&D and mass production in the field of millimeter-wave radar, Continental's radar products combine high performance and stability, and firmly occupy the leading position in the global and Chinese markets.
According to the Debang Research Institute, in China's millimeter-wave radar market in 2021, Continental's market share is as high as 23.9%, while in terms of forward radar, Continental's market share is 36.1%, which has been in the top three positions in the market for many years.
As the cornerstone sensor of intelligent driving solutions, millimeter-wave radar and individual sensors complement each other, while constantly iterating on their own. From the principle to the application on the car, from the engineering of the technology to the commercialization of the product to the cost reduction and efficiency increase of the product, millimeter-wave radar has gone through a long period of decades.
In order to get a share of this market, each player must have excellent full-stack R&D capabilities, system platform architecture capabilities, and experience in actual mass production projects.
In the field of millimeter-wave radar, a phenomenon that can be clearly seen is Tesla's use of 4D millimeter-wave radar, which has set off the pursuit of 4D millimeter-wave radar by capital, industry chain players, and vehicle manufacturers.
The world's first mass-produced 4D millimeter-wave radar was developed by Continental, and it is not difficult to foresee that a broad window is slowly opening for Continental.
Leifeng.com#Leifeng.com#Leifeng.com