The heat is soaring, and 4D digital imaging radar is opening the next wave of turbulence

2024-01-03

Millimeter wave radar has been deployed in vehicles for nearly 30 years, but its presence has not been strong, especially in recent years when the momentum has been suppressed by LiDAR. But the rapid iteration of 4D imaging capability and digitization, as well as the news of Tesla reinstalling 4D millimeter wave radar at the beginning of the year, have given millimeter wave radar a chance to take the lead again.

The actions of the capital market are the most intuitive direction. Since the beginning of this year, the capital enthusiasm for the domestic millimeter wave radar track has been soaring. According to incomplete statistics, at least 8 domestic 4D millimeter wave radar companies have received financing this year, with a total disclosed financing amount exceeding one billion yuan.

"Digital radar has enormous potential because it is crucial for road safety. Compared with existing analog radars, digital radar has a resolution improvement of 16 times, a target reflection power improvement of 24 times, and a contrast improvement of 30 times, providing road users with higher detection capabilities and road safety guarantees." Recently, at a technology forum held by Uhnder with the theme of "Digitization, New Trends", Manju Hegde, CEO and co-founder of Uhnder, spoke in an interview with GAC News.

4D digital imaging radar frequency output circle, L2+level replaceable LiDAR

Under the acceleration of automotive intelligence and the strong pressure of cost reduction and efficiency increase in the industry, the popularity of 4D millimeter wave radar is supported by product performance and cost advantages.

Compared to traditional 3D radars, 4D imaging radars have added detection of altitude (elevation) in the three dimensions of detection distance, velocity, and azimuth. They can track the motion trajectory of objects in real time and have better imaging capabilities.

According to the different signal modulation methods, 4D radar can be divided into analog radar and digital imaging radar. The former adopts an analog modulation frequency modulated continuous wave (FMCW) scheme, while the latter is based on a digital coding modulation (DCM) scheme. Although traditional radar and digital radar have their own advantages under different weather conditions, the upgrade of radar digitization will further enhance the hardware, software, and algorithm capabilities of 4D imaging radar.

Uhnder stated that generally speaking, 4D digital imaging radar performs better than traditional analog radar in terms of capturing static or dynamic object information on the road surface, resolution, contrast, target reflection power, anti-interference ability, sensor fusion ability, safety, and other aspects.

Thismainly due to the fact that traditional analog radar usually only has 12-16 virtual channels, with only one transmitter working at a time, resulting in low target reflection power. Traditional FMCW technology also does not have good resolution and anti-interference ability, and the market often requires about 5 sensor chips to support all functions.

As the pioneer of 4D digital imaging radar in the industry, Uhnder's 4D digital imaging radar solution, based on digital coding modulation technology, can provide up to 192 virtual channels, which is 12-16 times more than traditional analog radar. It supports 12 transmission channels and 2 x 8 reception channels, and can work simultaneously. Therefore, it can see a larger angle of view in the above dimensions and accurately process more targets, Provide a comprehensive improvement in the experience of autonomous driving.

In terms of cost, compared to the prices of LiDAR, which can range from five to six thousand or even tens of thousands, the thousand yuan cost of 4D imaging radar also holds an advantage. "Although the price of LiDAR is currently around $500, which has decreased compared to the previous $10000, the cost is still relatively high for mass-produced cars," Manju Hegde admitted.

According to Uhnder, traditional radar solutions often require 5, 7, or even 16 chips to achieve equal advantages, while Uhnder only requires one chip to achieve significant cost reduction and efficiency improvement.

"Our chip is embedded with signal processing design, so there is no need for additional processors, and the power is lower, about 1 GHz. It also integrates multiple Arm processors and a wide range of instructions, and all the software required for various radars is on the chip," said Manju Hegde.

It is worth noting that in the development of autonomous driving, the issue of which sensor technology is superior or inferior has always been a focus of debate in the industry. In Manju Hegde's view, in car cameras, radars, or LiDARs are all very important. With the upgrading of automotive intelligence, they are a cooperative and complementary relationship.

Manju Hegde pointed out that as autonomous driving advances, more precise perception needs to be provided. "For OEMs and Tier 1, the best thing is to be able to fuse all these sensors and achieve low-level fusion. If it is high-level fusion, it is likely that the disadvantage of one sensor will overshadow the advantage of another sensor. However, digital radar can relatively easily achieve fusion."

He further added that LiDAR has an advantage in optical frequency, and some good LiDARs can achieve micron level resolution. This high resolution is suitable for aircraft and satellites, but for current L2++levels and NOA functionality, a resolution of approximately 0.5 ° to 1 ° is sufficient, and 4D digital imaging radar can replace LiDAR. In the past, higher level development still requires the support of LiDAR. That is to say, there may be some competition between the two at the L2++level, L3 may achieve integration, and L4 must cooperate and complement each other. In addition, the integration of digital radar and camera can also better assist in improving the overall vehicle vision.

The acceleration of advanced autonomous driving will increase the market size of digital imaging radar

On November 17, 2023, the Ministry of Industry and Information Technology and other four ministries jointly issued a notice on the pilot work of intelligent connected vehicle access and road use, sounding a positive signal for the commercialization of L3; On December 5, the General Office of the Ministry of Transport issued the autonomous vehicle Transport Safety Service Guide (for Trial Implementation), which for the first time made it clear from the national policy level that ICVs can be used for transport operations.

It is not difficult to see that policy loosening is accelerating the arrival of the era of advanced autonomous driving. And this will undoubtedly further open up the market space for intelligent driving sensors.

Manju Hegde mentioned in her sharing that, based on the level of automation analysis, at the L2 level, the standard number of sensors per car is basically 3 radars, one front radar, and two rear radars. If collision warning or collision prevention system functions need to be implemented, two front radars need to be added on this basis, so there are a total of 5 onboard radars. So at L3, it may require 6 radars as standard, while L4 may need to reach 8-12 radars.

Although the entire 4D imaging radar market is still in its early stages of development, and most car models on the market also use traditional millimeter wave radar, the higher the level of autonomous driving, the higher the demand for safety, and traditional radar will not be able to meet more high-performance functions.

Uhnder also predicts that by 2025, the autonomous driving market will achieve the most basic safety functions, including AEB (Automatic Emergency Braking System) covering all low, medium, and high-end models, and then involving ACC (Adaptive Cruise Control System), LC safety functions, CTA, AES, etc., all of which require more advanced digital imaging radar.

Therefore, looking at market trends, radar is expected to experience explosive growth in the future, and the size of the 4D imaging radar market will continue to expand. According to Yole's statistics, the market size of 4D radar will reach 8.3 billion US dollars by 2028, an increase of 36% compared to 2022. Among them, the market size of 4D imaging radar is 2.2 billion US dollars, an increase of 49% compared to 2022.

It is worth mentioning that currently, 4D digital imaging radar is no longer just a theoretical technology, but has become a product that can be implemented and commercialized.

After Mobileye and Calterah, Uhnder is the third company in the world to achieve mass production of 4D digital imaging radar chips. It is understood that analog radars generally use 28nm technology, while Uhnder digital imaging radars have arrived at the 16nm technology node. It launched its first vehicle grade 4D digital imaging on-camera radar in March last year and completed relevant functional safety certification; Mass production began in July this year, with a current weekly production capacity of over 5000 pieces.

Given the importance of the Chinese automotive market and the advanced position of ADAS, Uhnder particularly emphasizes the "Chinese strategy". It is reported that Uhnder has established an office in China and recruited employees, and has communicated and exchanged ideas with multiple local Tier 1 companies. In the future, more resources will be invested to strengthen its development in the Chinese market.

At the end of the interview, when faced with the question of how Uhnder can better enter the market due to the emergence of some fixed sensor solutions for L2 level autonomous driving, ManjuHegde stated that Uhnder's 4D digital imaging radar solution can basically achieve seamless integration with existing solutions from Tier1 and OEMs of vehicle manufacturers. The actual deployment process is the same as traditional analog radar. "Our advantage lies in providing SPICE certified software design support, as well as APIs to facilitate Tier1 and OEMs' deployment of their algorithms on radar."

Top 10 IC design companies saw rapid revenue growth in Q3, and mobile phones and computers saw a rec