"Bosch has completely abandoned the development of self-driving LiDAR." I'm sure you're all aware of the news that the German newspaper Handelsblatt just broke.
Bosch abandoned LiDAR, because it invested heavily in research and development of MEMS (Micro-Electro-Mechanical-System, micro-electromechanical systems) technology complexity and long development time. A spokesperson for the company confirmed that "considering the complexity of the technology and the time to market, Bosch decided not long ago not to invest any additional resources in the hardware development of the LiDAR sensor."
However, many domestic media reports on this matter, but did not say is that, as the global automotive millimeter wave radar "ABCD" one of the four giants, Bosch hands there is "never give up" technology, that is, working in the 77Ghz band 4D millimeter wave Radar. This is where Bosch heavy protection.
Bosch also admitted that LiDAR is very important for L3 level autonomous driving. Because, although the industry 4D millimeter wave radar is taking off, but also can't "replace" the laser radar. However, 4D millimeter wave radar in the end what important potential, worth Bosch such a giant would rather give up the laser radar rather than give up?
It is worth mentioning that before this, another auto parts giant ZF has also withdrawn from the LiDAR competition. ZF had invested about 100 million dollars to participate in Ibeo (LiDAR "originator"), but with the bankruptcy of Ibeo, ZF also gave up LiDAR.
However, in February this year, Tesla submitted to the European regulatory agencies in the vehicle change application clearly, the upcoming production of the 4th generation Autopilot hardware HW4.0 added a high-resolution 4D millimeter wave radar. Previously, Tesla has removed the millimeter wave radar and ultrasonic radar, trying to All in visual route.
This kind of "hit the face" thing, the industry boiled, so you have a big eye also do this thing ah. Then, 4D millimeter wave radar suddenly became the industry's "new favorite". According to incomplete statistics, there are at least nearly 20 local enterprises to develop 4D millimeter wave radar products.
Why? The reason is that, compared with the traditional millimeter wave radar, 4D radar due to the increase in height information, can more accurately identify static objects, which is also known as the ability to have "imaging". Plus the price of about 1,000 yuan, compared to often six or seven thousand dollars or even tens of thousands of dollars of LiDAR, in the automotive industry "volume" to the extreme, the cost advantage is obvious.
And, compared to LiDAR, 4D millimeter wave radar inherited the traditional millimeter wave radar all-weather anti-interference advantages, and is not affected by light, smoke, dust, haze, in the night, rain, snow and other environments can work normally, adaptability is stronger.
Like CICC believes that 4D imaging radar can enhance the performance of millimeter-wave radar in all aspects, is expected to make millimeter-wave radar become one of the core sensors in the ADAS system, is an important direction for the future development of millimeter-wave radar.
In fact, 4D millimeter wave radar is not an unfamiliar new technology.At the end of 2018, in order to compete with Infineon and NXP, the two major manufacturers, Texas Instruments TI put forward the concept of 4D imaging millimeter wave radar, and launched a full set of 4-chip cascade 4D millimeter wave radar design solutions based on the AWR2243 FMCW (Frequency-Modulated Continuous Wave) single-chip transceiver, which integrates the antenna that is the most difficult to deal with for radar developers .
In March 2020, Google's Waymo released the fifth generation of automatic driving system perception program, millimeter-wave radar upgraded to 4D imaging radar, making 4D millimeter-wave radar technology for the first time applied to the vehicle end.
And in this year's CES speech, Intel's Mobileye CEO Amnon Shashua more emphasized the 4D imaging millimeter wave radar application scenarios in the car. He said, "by 2025, except in the front of the car, other places we only want (4D) millimeter wave radar, do not want LiDAR."
"Spring river water warm ducks know first", in fact, the domestic market 4D millimeter wave radar "volume" has long begun. Moreover, since the second half of 2022, 4D millimeter wave radar to accelerate the pace of "on board", the main carrying models are Feifan R7, Deep Blue SL03, Ideal L7 and so on. And some models are both laser radar, also want 4D millimeter wave radar.
Of course, as early as October 2021, more than half a year before, Bosch was in Shanghai for the first time to show the fifth generation of millimeter-wave radar Supreme Edition (as it was called at the time), with a maximum detection distance of 302 meters, a horizontal field of view of 120 degrees, and a vertical field of view of 24 degrees.
At the same time, Bosch's old rivals Continental, ZF, and Amberford are not idle. For example, Continental Group has the world's first 4D millimeter wave radar ARS540 put into mass production, ZF's 4D radar got the SAIC fixed point. The domestic market at the time, including Sensitech, Fritz Tektronix, for the ascension of science (CubTEK), Chuan speed microwave and many other companies have also entered the 4D track.
By 2023, the competition is even more white-hot, and more companies are investing in 4D millimeter wave radar research and development and mass production. Just on August 15, Li Yifan, CEO and founder of WoSai Technology, also participated in the first round of financing of Aoto Technology in his personal capacity.
It is reported that Altos V1, Aotu's first product to enter the mass production stage, has already done, "currently the world's only mature, non-FPGA 4-chip cascade (12TX, 16RX) 4D imaging radar products." And, in the comprehensive performance is not weaker than industry giants ZF, Bosch, Continental and other similar products performance, Altos V1 price is only 1/2 of the former or even lower.
Bosch can give up LiDAR, but will never give up 4D millimeter wave radar. Last August, Bosch and innovative RF antenna technology company from Sweden (GapWaves) reached an agreement to jointly develop and produce millimeter-wave radar antennas to meet the requirements of high-resolution in the car.
So, jokingly called "ABCD" (i.e., Autoliv, Bosch, Continetal, Dephi) giants, as well as the top domestic startups, have invested heavily, heavily in this, 4D millimeter-wave radar, the future of how, it is obvious.
Tesla used the "face" of the fact that even a pure vision program, still need 4D millimeter wave radar as a system redundancy. In addition, "multi-sensor fusion" as the industry generally recognized as a smart driving solution, 4D imaging radar will obviously occupy a place in it.
However, an important fact is that 4D millimeter-wave radar can not "replace" LiDAR, which is its technical and LiDAR heterogeneity decided. Of course, this does not affect the 4D millimeter wave radar bullish market.
4D millimeter wave radar benefits we do not say more, we only need to know clearly, 4D millimeter wave imaging radar research and development and application is not an overnight thing, there are still many technical difficulties need to continue to optimize and improve. From our table, we can also see that those who are able to mass produce 4D millimeter wave radar are still a minority, and most of them are still in the research and development stage.
Specifically, first of all, 4D radar needs more than one indicator at the same time to meet the requirements of the conditions of the host plant, the need to simultaneously improve the distance resolution, angular resolution, speed resolution, in order to achieve a better imaging effect, a single indicator of the power of the final imaging is not very meaningful, which is 4D millimeter-wave radar vendors, to do "door door excellent "is not an easy thing.
Secondly, to give full play to all the technical advantages of 4D millimeter-wave radar, it is necessary to do with the camera before the fusion, but, 4D millimeter-wave radar, the number of channels, the amount of data is relatively large, and the vision to do before the fusion of the arithmetic requirements of the higher, but also contains algorithmic issues, and the sensor side of the arithmetic is not enough, because of the millimeter-wave radar chip memory is limited.
Therefore, if the point cloud density of the data is relatively high, the pre-fusion needs to be done in the domain controller. But this also has problems, on the one hand, the high data rate and data compression of 4D millimeter wave radar will bring challenges to the centralized architecture, on the other hand, the bandwidth and rate of signal transmission between the antenna and the processor will also affect the detection accuracy.
Therefore, to solve these contradictions, 4D millimeter wave radar vendors need to have a deep enough understanding of the central domain controller, or with a domain controller vendors or chip vendors deep binding. For example, Ambarella Semiconductor (Ambarella) acquired Oculii Radar (Oculii), launched a new generation of 4D imaging radar. But we know that these core technologies at least few domestic companies can control at present.
In addition, the former fusion needs 4D millimeter wave radar with the camera to do joint calibration, but the joint calibration is difficult. In addition, 4D millimeter wave radar has distance information, while the camera does not. Then how to deal with the two when they are jointly calibrated is also a big problem.
Moreover, the problem of the number of antenna channels and the chip cascade also has strong technical content. For example, some domestic companies mostly refer to the traditional design experience of ABCD, and 4 chip cascade 12T16R (12 transmitter channels, 16 receiver channels) antenna array if their own design from scratch, just the amount of simulation and server requirements have constituted a threshold.
Another problem with 4D millimeter wave radar is that EMC (electronic compatibility) is difficult to pass. The key to this is the need to consider how to avoid interference with the outside world (including interference with objects outside the vehicle as well as the car radio inside the vehicle, etc.) and how to combat interference from the outside world. However, EMC issues are difficult to detect in the pre-simulation and need to be discovered during the experiment.
Compared to LiDAR, 4D millimeter wave radar is considered cost-effective. However, compared with the mature 3D millimeter wave radar, its advantage is not so obvious at present. Therefore, 4D millimeter wave radar can only be regarded as "warm-up", some industry insiders believe that it will take 2-4 years to mature.
Of course, some people are very optimistic. Chu Hang Technology CEO Chu Wing Yan said in an interview with the media, "millimeter wave radar is a strategic point of automotive perception, is the only sensor with all-weather sensing capabilities. It has the richest amount of perceptual feedback, distance, speed, angle, horizontal angle of everything, now 4D, the future may do 5D, 6D, 7D."
However, Chu Wing Yan also admitted that 4D millimeter wave imaging radar mass production still needs to solve some problems, such as testing. This includes the need for sufficiently accurate test equipment to test whether the radar system meets customer requirements. However, the industry currently lacks a unified standard evaluation method and means.
5D, 6D, 7D or whatever is still far away, but 4D millimeter wave radar still has potential. Data from the GaoGong Intelligent Vehicle Research Institute predicts that the millimeter wave radar carrying rate of new cars in the domestic market L2 plus and above is expected to exceed 50 percent in 2025.
At the same time, 4D millimeter wave radar is expected to 2023 carrying volume has the opportunity to exceed one million, to 2025 accounted for the proportion of all forward millimeter wave radar is expected to exceed 40 percent .
