Shulou(Shulou.com)11/24 Report--

It is well known in the industry that lidar is rejected by Musk.

"people who use lidar are idiots, like people with a bunch of appendices." Musk knocked over a boatload of people at the "Autonomy Day" conference in 2019. Later, Musk judged again that "those who rely on lidar are doomed to failure."

Pure visual perception scheme is the technical route that Tesla has always adhered to. In 2021, Tesla became more radical. In May of that year, he announced that he would take off the millimeter wave radar on the newly produced Model 3 / Y in the North American market, causing a lot of controversy in the industry. Recently, Tesla cut off the last radar again, and Musk finally walked on his ideal road.

However, as loyal fans of Tesla, Weilai, Xiaopeng and ideal frequently make friends with lidar. In 2021, Xiaopeng as "the first person to eat crabs", Xiaopeng P5 was the first to load two lidars. Subsequently, many car companies followed, opening the road of "lidar getting on the bus". Accompanied by a number of Chinese radar enterprises gradually come to the fore, from the global scope, investment and manufacturing are beginning to tilt to China. Most established radar companies did not expect that the first stop on the train would be China.

In a short period of more than ten years from 2004 to the present, lidar has experienced various twists and turns, brilliance, loss and landing in China, which is related to the national conditions of China and the United States, and to the situation of traffic roads. the most important thing is related to technology and commercial logic and timely selection of tuyere. In this article, the car think Tank will mainly try to answer the following three questions:

1. In the United States is not as expected, in China "getting on the train", why is the development of lidar so different at home and abroad?

two。 How did the lidar "boarding tide" rise in China?

3. What are the opportunities and challenges faced by domestic lidar enterprises?

Rejected by foreign markets: the giant went bankrupt in 2004, the Defense Advanced Research projects Agency (DARPA) launched a self-driving challenge (DARPA Grand Challenge), which was the starting point of lidar.

There are three competitions. Between 2004 and 2007, nearly 20 teams completed the competition, including the application of lidar-related technology. Behind the technical support is the famous radar company Velodyne and its founder David Hall.

In 2009, Velodyne began to officially sell the world-famous mechanical rotary lidar ancestor HDL-64E. Its core concept is "time of light flight" (Time of Flight, referred to as ToF): a laser emits a laser, which bounces back when it touches an obstacle, and the speed of light is known, so the distance between the emitter and the obstacle can be calculated by recording the time when the light goes back and forth. If the laser beams emitted by the emitter are enough and dense enough, these points come together to form an image, which is called a "point cloud" image.

The bigger change brought about by the DARPA Challenge is the prelude to the commercialization of high-level autopilot. Another big-name spectator watching the DARPA challenge was Google founder Larry Page, who later invited Sebastian Thrun, a Stanford AI lab professor who represented Stanford, to join Google and set up an autopilot division.

Over the next few years, Cruise, Zoox, Argo, Nuro. A large number of American self-driving start-ups were born in Silicon Valley. Google's self-driving division also officially became an independent business entity at the end of 2016, later known as Waymo. Later, at one point, the market capitalization of Waymo soared to $175 billion, equivalent to one of the four GOOG companies.

In the near time, some Chinese scientific research institutions and companies have also begun to develop autopilot technology. In 2011, the University of National Defense Science and Technology completed a self-driving test of 286 kilometers from Changsha to Wuhan for the first time, and in 2013, Baidu's first self-driving car based on BYD Qin was unveiled at Wuzhen Congress.

Perception, planning and control-these are the three main parts of the autopilot system, which is analogous to human driving, that is, the eye can see the road, the brain knows how to drive, and the hands and feet can control the car. Lidar solves the problem of "perception", which is the "eye" of high-level self-driving vehicles.

Car companies hope to make cars "far-sighted" through the ranging function of lidar, which can be "seen clearly" even in rain and snow. Lidar is mainly divided into mechanical, semi-solid and pure solid, which is also the three stages of lidar development. Radar has always been regarded as the main component above L3 in the field of autopilot, and its main function is to measure the distance, and at the same time, it is not affected by the weather such as rain and snow, and is mainly used to make up for the lack of camera "car eyes".

Of course, some people think that the "car eye" itself can complete intelligent assisted driving. Take Tesla as an example, the reason why Tesla removed the radar one after another is that the cost is too high. At present, the price of a high-performance lidar can reach tens of thousands of yuan, and the cheap one costs 3000 yuan, while the price of a high-definition car camera module is only 400 yuan, and the estimated cost of the FSD chip developed by Tesla is only 200 US dollars.

As the oldest lidar company in the United States, Velodyne's radar never boarded Tesla's car. "in the second quarter, we found strong customer attraction in three target markets, industrial and robotics, smart infrastructure and self-driving cars," Ted Tewksbury, chief executive of Velodyne, said at the second-quarter earnings report. Its main customers include Boston Dynamics, the global leader in the field of mobile robots, and the traffic safety improvement project in Helsinki, Finland. In fact, it has something to do with an early decision and market judgment made by Velodyne.

Back in 2015, many lidar companies have seen that another major customer base is car companies.

Car companies believe that they need to acquire self-driving ability to catch up with the trend of intelligence. Instead of doing high-level self-driving directly through non-commercial road tests with strict qualifications, car companies dominate another progressive autopilot route: starting from the auxiliary driving of L2 and L3. Sell the cars with these functions first, and then improve the self-driving ability with the help of a large number of real data optimization systems generated by users in the process of driving. Although the progress is slow, this approach is of great benefit to mass production.

But Velodyne decided in 2019 that Chinese car companies were not lidar beginners, and Velodyne chose to withdraw from China. The next two events hit Velodyne in the face:

First, new car-building forces such as Ulay, Xiaopeng and ideal have been born in the Chinese automobile market around 2015, and the high-end brands also hope to highlight their intelligence and scientific and technological attributes by using lidar, which is different from Tesla, who insists on not using radar.

Second, it is generally believed in the industry that solid-state and hybrid solid-state technology is the future of the next generation radar. When the competitors who had been left behind changed lanes to outflank the hybrid solid-state route, Velodyne failed to continue to come up with leading products.

These two things also put Velodyne in a completely passive position. This means that Velodyne missed its nearest opportunity for mass production to reduce costs and had to rely on capital and the secondary market to supply high R & D spending. Recently, Velodyne had to merge with Ouster, another listed lidar company, in order to survive.

Of course, the opportunity will not wait for anyone, in the subsequent competition, the advantage came to several Chinese enterprises.

It is sought after by Wei Xiaoli: in the first year of getting on the bus in China, lidar received a complete "courtesy", in sharp contrast to the collapse of companies in the United States and survival together. Lidar mass production has become a natural thing in China.

According to car Intelligence Tank, there are 21 lidar models and 16 Chinese companies that have been mass-produced / planned to be mass-produced in 2022. Among them, the Xiaopeng P5 carries 2 lidars, 5 millimeter wave radars and 12 ultrasonic radars; the ideal L9 has a lidar with 128 lasers on the top of the car, and the Weilai ET7 model is equipped with a lidar with a maximum detection distance of 500m. In the self-driving system of the Lulai car, the lidar and 11 cameras have the highest priority.

There are signs that Chinese enterprises are moving towards the forefront of lidar. According to the 2022 Lidar Application report in Automotive and Industry released by Yole Intelligence in August this year, Wesai Technology ranked first in total revenue, followed by Velodyne and Ouster in the global automotive lidar field. From this ranking point of view, relying on the vast domestic new energy vehicle market, Wesai Technology and Suiteng Juchuang have indeed established their own advantages.

The inflection point that impels the lidar to get on the bus is the landing of the urban navigation assisted driving (NOA).

When a car company plans a hardware to get on the car, it all has the corresponding function definition. Judging from the functions of intelligent driving announced by various car companies, urban NOA is the main use scene of lidar. Lidar also gets the opportunity to get on the bus with the landing of the city NOA. The main reason is that car owners use cars more frequently in cities, so the navigation assistance value of urban scenes is higher than that of high-speed navigation. However, in the urban scene, the roads are complex, vehicles are dense, and emergencies occur frequently.

Safety has become a higher priority issue, and Tesla's several "ghost brake" incidents are undoubtedly telling the industry that multi-sensor fusion programs are more secure. When more car companies want to take the single intelligent driving route of "emphasizing perception and neglecting map", the importance of lidar has been vaguely elevated again.

Recently, Wo Sai Technology and Suiteng create two "senior players", as agreed in general, have released their own lidar. In August, Suiteng Juchuang released the all-solid-state blind lidar RS-LiDAR-E1;11 month, and Wesai Technology officially released the pure solid-state lidar FT120 for ADAS front-loaded production vehicles. This is another breakthrough after the mass production of the widely acclaimed long-range hybrid solid-state lidar AT 128.

For car companies, a more important problem is how to make good use of lidar. According to the report of "late Latepost", the current mainstream scheme is changing to "image pre-fusion", that is, the front-end fusion is done before the point cloud and image data are calibrated independently, and then go to the algorithm layer for recognition. This also means that lidar is more closely related to chips and intelligence while driving, and it is also more difficult to "give up".

Since 2020, there has been a sudden rush of investment in radar companies among car companies: BYD has invested in Suiteng, Weilai has invested in Tudatong, Xiaopeng has cast a path, and Xiaomi has invested in Wo Race. According to Che Baixu, more Chinese car companies are considering investing in lidar companies, especially second-tier companies with relatively low valuations. The move is seen as a way for car companies to lay out supplies in advance to prevent "overheads" in radar companies from causing prices to be too high.

The ultimate problem: cost-a problem that bothers not only car companies, but also radar companies.

Why does the lidar company keep getting bad news? There are only two words that trap them: cost and getting on the bus.

Velodyne started as a mechanical lidar, which is relatively accurate but bulky and more expensive. According to titanium media, when Velodyne first launched the HDL-64E in 2007, its price was as high as $80,000, equivalent to the price of an ordinary car, and remained so high until 2014.

Nowadays, mechanical radar has been abandoned by most vehicle manufacturers because of its short life and high price, which is mainly used in the field of self-driving, while the solid-state lidar, which is mainly based on MEMS and rotating mirror scheme, is not only cheaper for on-board front parts, but also easier to reach the standard of vehicle specification. At the same time, the MEMS lidar, which can be integrated into the car body, can make the whole vehicle more beautiful than the mechanical lidar that needs to be placed on the roof.

From the perspective of market price, the cost of solid-state lidar is still the main limiting factor for its development. It is reported that when Suteng Juchuang 125-line solid-state lidar RS-LiDAR-M1 Simple went on sale in January 2020, the unit price was $1898, while the current price of Wesai AT128 radar is about $700. By contrast, the price of a high-definition car camera module is only 400 yuan.

The wide price gap has also deterred many car factories, including Tesla. Ranging has always been one of the abilities that can not be ignored in navigation-assisted driving. Yang Yongcheng, a partner at Fengrui Capital, once told the media: "if lidar becomes cheaper, manufacturers will naturally compromise with users and the market."

At present, several domestic head radar enterprises have been bound with car companies to open mass production, and Suiteng Juchuang binding mainframe factory has harvested the fixed points of nearly 50 models. Wesai Technology has also obtained a fixed point of mass production of millions of models from many car companies, such as ideal, Changan, Jidu, Gaohe, Lutes and so on. Among them, Wesai Technology's AT128 radar has delivered more than 10,000 in September 2022, making it the first vehicle lidar company in the world to deliver more than 10,000 a month.

For lidar, there is an urgent need to solve the cost problem. Cost reduction depends to a large extent on mass production, but only by knocking down the cost can it be recognized by the car companies.

This is a good time for Chinese lidar, and Chinese enterprises are at the forefront of the world in terms of technology and market. "overseas radar enterprises started too early, but due to the delay in mass production due to the unsolved key technical problems, the financial statements are not good-looking, so there are all kinds of problems." "in the early days, Velodyne was very expensive, but when the car companies gave them orders, they were unable to deliver, and now they have withdrawn from the Chinese market," said Zhang Xiang, a special expert at the CBL think-tank. "now Chinese companies are conquering all the key lidar technologies."

At present, it is an important moment for China's new energy vehicle enterprises to overtake at the bend. If the cost problem can be solved, radar will become an important advantage of overtaking in Chinese corners.

[reference]

[1] "misjudgment, counterattack, overturn, lidar ten years on the road" Zhang Licheng Cheng Manqi, Latepost later

[2] "Lidar, there are a lot of plays." Zhang Zhidong, C dimensional

[3] "Lidar industry": foreign companies "spend the winter together", domestic enterprises "roll" to the new track "Sun Lei, Daily Economic News"

[4] Musk said, "those who use lidar are fools", but the Chinese don't believe Pan Yuchen, Observer Network.

[5] "merger, delisting, bankruptcy, the cold winter of lidar PLAYERUNKNOWN'S BATTLEGROUNDS" Han Jingxian, titanium media

[6] "lidar hustle and bustle" bell, red interstellar

[7] "there is no room for a small radar for such a big Tesla? Xiong Yuxiang, Yuchuan Automobile Review

This article comes from the official account of Wechat: che Bai think Tank (ID:EV100_Plus), author: Yang Deze

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.