The cost is much lower than that of LiDAR. Is there a new helper for advanced autonomous driving?

Recently, Shanghai Geometry Partner Intelligent Driving Co., Ltd., a provider of overall solutions for autonomous driving, announced the completion of a new round of strategic financing. Since its inception, the company has chosen to use 4D imaging millimeter-wave radar as its main sensor and combined it with other perception systems to develop L2~L4 autonomous driving solutions.

In fact, there are more than one company targeting 4D imaging millimeter-wave radar. At the 2022 International Consumer Electronics Show (CES), a number of chip companies such as NXP, TI, and Mobileye launched or updated their own imaging radar solutions, and millimeter-wave radar system manufacturers such as Arbe and Zadar Labs also brought their own imaging radar products. As for domestic manufacturers, technology companies such as Continental, Aptiv, and Huawei have also launched 4D millimeter-wave radar products, and a number of emerging players such as Aoku, Suzhou Millimeter Wave, and Anzhijie have also begun to deploy in this field.

Why has 4D imaging millimeter-wave radar become the "new favorite" in the autonomous driving industry? Compared with previous perception radars, what are its advantages?

Make up for the shortcomings of traditional millimeter-wave radar performance

4D imaging millimeter-wave radar can be seen as an upgraded version of the previous millimeter-wave radar.

It is understood that traditional millimeter-wave radar determines the relative distance, speed, angle and direction between the car body and other objects by emitting a conical electromagnetic wave beam and detecting the echo, and then submits the decision to the central processing unit.

In terms of detection range, traditional millimeter-wave radar can only detect three dimensions: distance, direction, and speed. Due to the lack of ability to measure height, low angular resolution, sparse point clouds, and ignoring static objects, traditional millimeter-wave radar is still acceptable in autonomous driving assistance systems below Level 3, but it is incapable of higher-level autonomous driving.

The 4D imaging millimeter wave radar adds data analysis of the target height dimension on the basis of the original distance, direction and speed, which can realize information perception of the four dimensions of "3D + speed", and can effectively analyze the outline, category and behavior of the target, and depict a more realistic environmental image. Relying on the height measurement capability and point cloud image, the 4D imaging millimeter wave radar can preliminarily determine the positional relationship between the stationary object and the vehicle, avoiding the safety hazards caused by simply filtering the stationary signal.

Moreover, the price of 4D imaging millimeter-wave radar is between 100 and 150 US dollars, and the mass production cost in the short and medium term is much lower than that of laser radar. "It will take at least five years to reduce the cost of laser radar from several thousand yuan to several hundred yuan, so 4D imaging millimeter-wave radar can replace low-beam laser radar to some extent," said Xie Jianjun, president of Aoku Radar Asia Pacific.

Some people believe that, considering the comprehensive performance and cost, 4D imaging millimeter-wave radar will quickly occupy an important position in scenarios such as L3 autonomous driving and unmanned autonomous parking. Because these scenarios often pursue the ultimate cost-effectiveness and cannot adopt lidar on a large scale, but they also need some of the advantages of lidar such as stationary target detection, which are precisely the advantages of 4D imaging millimeter-wave radar.

CICC believes that 4D imaging radar improves the performance of millimeter-wave radar in all aspects, and is expected to make millimeter-wave radar one of the core sensors in ADAS systems, which is an important direction for the future development of millimeter-wave radar. It is expected that by 2025, the market size of China's automotive 4D imaging radar will reach US$190 million, US$360 million and US$540 million respectively under pessimistic, neutral and optimistic scenarios, and the annual compound growth rate from 2022 to 2025 may be 34%, 64% and 88%.

Large-scale production still needs time

Although the overall cost-effectiveness of 4D imaging millimeter-wave radar is high, there is currently no product that can be truly put into large-scale implementation.

"As an automotive-grade safety component, the 4D imaging millimeter-wave radar has a very long development cycle, and sufficient time is required to complete system design, testing and verification. Generally speaking, it takes five years to develop a radar. If you go the customized chip route, it may take even longer," said Zhang Wogong, chief technology officer of Chuhang Technology.

At the same time, there is currently no test equipment for 4D imaging millimeter-wave radar in the industry. In other words, the industry ecosystem is not yet mature. Manufacturers mostly use traditional millimeter-wave radar test equipment to verify the performance of 4D imaging millimeter-wave radar products. However, traditional test equipment has a low resolution and cannot verify whether the resolution of 4D imaging millimeter-wave radar has reached 1 degree or less, which is crucial for the radar to distinguish objects such as cars and people at a relatively close distance.

"We can only go on the road and use a truth value system and lidar for comparison. But in this case, if there is a problem with some basic performance of the product, such as antenna design, we have no way to detect it in the laboratory, we cannot do front-end testing, and we cannot form a complete R&D closed loop." Chu Yongyan, CEO of Chuhang Technology, predicts that it will take another 3 to 5 years for 4D imaging millimeter-wave radar to become a standardized, mass-produced product.

In addition, from a technical perspective, in a complex electromagnetic environment, millimeter-wave radars are subject to greater interference from noise and other sources. In a 4D imaging radar that pursues precision, how to deal with noise and possible false warning noise is also a difficulty in the application of 4D imaging radars. As more and more vehicles use millimeter-wave radars, mutual interference between millimeter-wave radars will also increase, and anti-interference will also be a problem that subsequent 4D imaging radars must face.

"This year we have seen several OEMs preparing to install 4D imaging millimeter-wave radars in their vehicles. This is just like the lidars a few years ago. Once a leading company applies it on a large scale, other companies will quickly follow suit," said Xie Jianjun.

However, some industry insiders believe that 4D millimeter-wave radar is only a part of the automotive autonomous driving sensor system, including optical sensors. In the future, it may be combined with lidar and cameras to promote driving safety at a lower cost and help the commercialization of autonomous driving technology.