In the field of intelligent driving, the mass production of many new technologies is later than expected, but lidar is an exception.
Entering 2021, under the leadership of Weilai, Xiaopeng, Ideal, BAIC Jihu, GAC Aian, Great Wall and other brands, more and more car companies have begun to announce the installation of lidar on new cars, and the number of bicycles is also increasing. For example, the first model of Avita Vita 11, the new high-end intelligent electric vehicle brand of Changan Automobile, is equipped with 3 laser radars, and the first model of the Salon brand newly launched by Great Wall Motors, the Mech Dragon, even has reached 4 laser radars.
In terms of delivery time, many of these models are released this year and will not be officially delivered on a large scale until next year. According to incomplete statistics from Gesta Motors, at least 15 new cars equipped with lidar will be listed next year. This means that 2022 is expected to truly become a big year for the mass production of lidar.
So, considering the performance and cost of lidar and other factors, is it time for large-scale mass production of this technology to be put on the car? In this regard, Gaz Automobile recently launched an industry survey, the results show that in this survey of hundreds of people, only 30% of the participants believe that the current time for lidar to "get on the car" is ripe, and the vast majority of them believe that the time is not really ripe, especially the performance, cost and market basis of lidar, etc., which need to be further optimized.
Image source: Gaz Cars
When lidar got on the bus, only 30% of the participants thought the time was ripe
For the past few years, lidar has been considered the standard for L3+ autonomous driving. In the L1 and L2 stages, although the driver can be allowed to properly feel off (foot release) and hand off (hand release), the driver's eyes need to keep an eye on the road conditions. In this case, even if there are certain limitations in millimeter-wave radar, cameras, etc., they can still meet the needs.
But in L3, according to the current definition, drivers are allowed to remain eyes off in certain scenarios. For example, Mercedes-Benz has just passed the L3 automatic driving system approved by the strict technical regulations of the German Federal Motor Transport Administration (KBA), which allows the intelligent pilot system to take over the driving task in the high-speed section with suitable conditions and dense traffic, and the driver can perform other operations, such as contacting colleagues through the on-board mobile office system, surfing the Internet or watching movies.
This means that in the L3 stage, the sensor configuration on the vehicle can completely replace the eyes of the human driver in some scenarios, which requires an exponential increase in the requirements of the perception system compared to L2 or even L1. Therefore, lidar with higher perception accuracy and longer detection distance is taken into account, on the one hand, to make up for the defects of millimeter-wave radar, cameras, etc., on the other hand, to improve the redundancy of the perception system.
Guo Jishun, vice president of Joyson Electronics and president of joyson intelligent vehicle technology research institute, has previously said in public that he is a staunch supporter of lidar. "I think only a fully heterogeneous sensor network can form a robust system."
However, this year, although L3+ automatic driving has not entered the reality as scheduled, many car companies have announced lidar mass production plans. In this regard, 26% of the participants believe that the rapid evolution of the automatic driving function of the main series production vehicle needs the support of lidar.
By combing through the planning of major car companies for lidar mass production models, it is not difficult to find that many of these models are directly running to L3 in the design of automatic driving functions, or are infinitely close to L3 in experience. For example, Xiaopeng P5, based on the new XPILOID 3.5 automatic driving assistance system, can further realize urban NGP on the basis of the high-speed NGP that has been achieved so far. The nippon ET5, which has just been launched, will gradually realize the point-to-point automatic driving experience of high-speed, urban area, parking, power exchange and other scenarios such as high-speed, urban area, parking, and power exchange through NIO's latest autonomous driving technology NAD, as well as NIO's supersensitive system Aquila and supercomputing platform Adam.
These are the typical application scenarios of L3 autonomous driving planned by OEMs at this stage. Although L3 autonomous driving still faces multiple challenges in terms of technology, responsibility division, and supervision, it has been confirmed that it is expected to take the lead in landing automatic driving in special scenarios such as highways, urban expressways, and traffic congestion. And many car companies have been technically achievable, only to wait for the real "ice-breaking" of the policy.
From this point of view, at this stage, many car companies announced that lidar is on the car, which is also based on a hardware embedding strategy. That is, regardless of whether the time for high-end automatic driving mass production is ripe now, the number of lidar, cameras, etc. is piled up first, so that L3 and L3+ autonomous driving can be realized through OTAs at the right time in the later stage, or when the policy is liberalized. Wei Bofan, director of bmw group in charge of research and development, previously said that in the future, BMW iX will be able to obtain L3 level autonomous driving system through software updates, because iX has technically met L3 functions.
Great Wall's coffee intelligence also supports OTA upgrading L3 to L4 and even higher level of autonomous driving. It is not difficult to understand why WEY Mocha boldly equipped with 3 solid-state lidars, in addition to equipped with up to 8 millimeter wave radar, 8 cameras and 12 ultrasonic radars, this car was running to L3 at the beginning, and even higher level of automatic driving.
In addition to the needs of technological evolution, 20% of the participants believe that the current time for lidar to get on the car is ripe because of the brand upgrade and product positioning needs of independent car companies.
Especially in the past two years, the brand upward has become the main theme of the development of independent car companies, and in the car companies with lidar mass production plans, in addition to luxury brands such as Mercedes-Benz, BMW and new car-making forces such as Wei Xiaoli, there are indeed many high-end brands of independent car companies, such as BAIC Jihu, GAC Aean, Changan Avita, SAIC Feifan Automobile, Great Wall WEY, Great Wall Salon Automobile, etc. Although Geely's high-end brand Krypton Automobile has not yet announced its lidar mass production plan, it has also recently reached a cooperation with Sagitar Juchuang, and does not rule out that it will be equipped with Sagitar Juchuang's lidar on relevant models in the future.
Image source: Great Wall Motors
For these high-end brands, intelligent driving, especially high-level automatic driving, is undoubtedly an important gene, especially when everyone begins to work on this track, and if you want to stand out, you must have a better performance in technology. At this time, as a symbol of advanced automatic driving, lidar has naturally become everyone's first choice.
And it is undeniable that with lidar, whether in the capital market or the mass production car market, automakers also have better storytelling materials.
Performance, cost, etc. are still constraints, and most people think that the time for mass production is not ripe
Admittedly, the evolution of autonomous driving is inseparable from the support of lidar. But in the eyes of many people in the industry, it has not yet reached the point where lidar is indispensable. According to the survey results, 47% of the participants believe that car companies have released lidar mass production plans this year, but the time is not yet ripe.
Behind this, the performance and price of lidar can not meet the requirements of the mass production vehicle market, and the evolution of high-end automatic driving is still a large uncertainty are the two main factors, which is also the key basis for large-scale mass production of lidar.
Taking the price as an example, according to this survey, if the lidar is to be mass-produced, 94% of the participants accept the price below 5,000 yuan, of which 25% believe that the price is controlled below 500 yuan, 39% between 500-1000 yuan, and 30% between 1000-5000 yuan. Obviously, the vast majority of lidar is not yet able to meet this demand.
According to relevant analysis data, the average price of lidar that can be mass-produced is about 1,000 US dollars, equivalent to about 6,000 yuan. If the batch is loaded, considering the large-scale cost reduction, the cost may be slightly reduced. Taking the Xiaopeng P5 as an example, the 550E model without lidar is priced at 192,900 yuan, while the 550P model equipped with two lidars is priced at 199,900 yuan, with a difference of 7,000 yuan in the middle, which means that the price of a single lidar is about 3,500 yuan.
Image source: Xiaopeng Motors
But if it's going to fall any further to $100 or $200, it's going to be a few years in the future. Huawei has previously said it plans to reduce the cost of lidar to $200 (about 1,276 yuan) or even $100 (about 638 yuan), but Huawei did not give a clear timeline. It is reported that the cost of its 96-line hybrid solid-state lidar is to control the target within $500.
Tutton co-founder and CEO Junwei Bao expects that when Tutton's annual shipments are 100,000 units, the cost will drop to about $1,000. The NIO ET7, which will currently be equipped with Tuduitone lidar, has not yet begun to be delivered, which means that this target of 100,000 units will not be easy to achieve in the short term.
Also included Sagitar Juchuang has also disclosed that if the order size reaches 100,000-1 million units, the hardware price can be lowered to 200-500 US dollars. According to a previous report by Yole Développement, sagittarius Juchuang (10%) of the global ADAS market currently has a market share second after Valeo (28%), ranking second. Although the actual shipments of Sagitar Juchuang are not yet known, according to Valeo's latest statistics, it has produced more than 150,000 lidars, which means that Sagitar Juchuang still has a long way to go to achieve the above goals. In the short term, cost will continue to be a major constraint on lidar.
The other is the evolution of autonomous driving, which does not provide a good market basis for the mass production of lidar. As mentioned earlier, the most important stage of lidar is L3 and higher automatic driving, but at present, except for a few markets such as Germany and Japan, which are trying to release L3, the vast majority of markets do not allow L3 self-driving cars on the road, including China.
"Although there are now some leading car companies that claim to have reached the L3 or even L4 level of automatic driving, but because of different local regulations and policies, the current automatic driving level certification still lacks a unified official certification." In terms of function, most of the models of automatic driving function is still stuck on L2 or L3, and there is still a long way to go before full automatic driving. Recently, Zhu Xiaoyue, Business Development Manager of Greater China of Keysight Automotive Electronics and Energy Division, said at the media communication meeting.
The inability of high-level self-driving cars to hit the road means that the practical utility of lidar will be greatly reduced. Even in the L2 and L2+ automatic driving stages, lidar can also be used to improve perceptual safety, but the question is whether such an expensive sensor is needed in the assisted driving stage, in fact, it is debatable, after all, the wool is ultimately out of the sheep.
Moreover, at this stage, lidar is also facing the problem of lacking the support of corresponding evaluation standards. In particular, the performance requirements for mass production vehicles on lidar, as well as the environmental test conditions of vehicle regulations, have not yet formed standards in the industry, which leads to the lack of transparency in the current integration industry.
Image source: Valeo
At present, when lidar manufacturers describe product performance, they often stay at the farthest detection distance of standard targets, but many of them are actually the perception results in the ideal environment, and in the actual application process, such as rain and fog, sunny backlight, haze, etc. will become interference factors in the actual use of lidar. And according to the different application scenarios of the auxiliary driving function, the demands of automakers in the process of lidar selection are actually different. To this end, in October 2021, the Electronic and Electromagnetic Compatibility Sub-Technical Committee of the National Automotive Standardization Technical Committee has organized the drafting of the standard "Performance Requirements and Test Methods for Vehicle-Mounted Lidar", but it will take some time to form a formal standard.
Another is the after-sales problem, at present, lidar such a high-tech technology can not be like the general auto parts for the time being, there is a problem directly to the 4S shop or other maintenance points for maintenance, which will actually bring great inconvenience to the later use of the car, after all, once the lidar can not be used normally, it means that other related automatic driving functions will also be affected.
From this point of view, lidar has improved greatly compared with the previous two years in terms of product performance and price, but the real "getting on the car" or not actually requires many considerations from the automaker, such as product positioning, practical scene application and whether the follow-up market and policy basis can give sufficient support. For automakers, following the trend of mass production of lidar does not mean that the product performance is necessarily excellent, on the contrary, it does not carry lidar, nor does it mean that the product is not outstanding, in the final analysis, it is best for target consumers to truly create user value.
At present, automakers actively promote the lidar on the car, although the release of favorable signals in the industry, combined with various factors, it is expected that it will take a certain amount of time to cultivate the market. Although there may be a big leap in the number of lidar markets next year, the real large-scale mass production may start from 2023 and gradually increase in the next few years to move towards the outbreak period.