"Science and Technology Innovation Board Daily" (Shanghai, researcher Song Ziqiao) reported that after "rebellion" against Bitcoin, Musk "changed his face" again. Different from his previous clear-cut opposition to the use of lidar, he It is finally ready to be used in Tesla’s autonomous driving technology testing.
According to people familiar with the matter, Tesla signed a contract with lidar supplier Luminar for testing and development using laser sensor technology. As part of the agreement, Luminar sells sensors to Tesla for use in the Model Y.
Tesla has always claimed that it is not optimistic about lidar. Musk bluntly said at Tesla's first "Autonomy Day" in 2019, "Only fools will use lidar. We are optimistic about laser." All radar companies will be doomed." Last year, he even stated that “Lidar will not be used even if it is free,” and he wanted to algorithmically solve the problem of visual information processing.
Musk, who is so stubborn, finally extends an olive branch to lidar. Why now?
Pursuing cost-effective products
Cost-effectiveness may be the real reason why Musk decisively abandoned lidar in the first place.
Omer Keilaf, CEO of lidar startup Innoviz, said in an interview with foreign media at the end of February this year, “Five years ago, when he (Musk) had to decide to bring his car to the market At that time, there was no lidar with suitable price and mature performance, so he made a good decision (not using lidar, but focusing on cameras and radar). ”
But Keilaf believes that the situation has changed now. Different from the past. “LiDAR components that previously cost tens of thousands of dollars now cost just $1,000, and will soon be down to $500, and down to $100 for the foreseeable future. And startups like Innoviz will soon be valued at $1.4 billion SPAC reverse merger to go public," Keilaf said.
Huawei’s latest lidar is as low as $200.
In addition, at that time, most lidar structures were still relatively rudimentary and had disadvantages in terms of stability and service life. Su Shuping, general manager of Innoviz China, said in an interview in June 2020, “From the information I know (which may not be comprehensive), the vast majority of lidars currently on the market can work for 8 or 4 hours a day. , The lidar used in the last year is of very good quality.”
Now, lidar technology has experienced an evolution from mechanical to solid-state in the field of autonomous driving. In 2021, many companies have launched it. Car companies have launched new models equipped with lidar and announced mass production plans, which brings development opportunities to the lidar industry.
Recently, Pony.ai and lidar manufacturer Luminarl jointly released an integrated autonomous driving sensing system; ArgoAI, an American autonomous vehicle technology company funded by Ford and Volkswagen, recently developed a long-range sensing system A new lidar system up to 400 meters. During this year’s Shanghai Auto Show, ARCFOX Alpha S, Xpeng P5, NIO ET7, SAIC RS33, BMW iX, and Zhiji L7 all announced the launch of lidar.
Frequent safety accidents may encourage Tesla to adopt lidar
Currently, doubts about the safety of Tesla’s autonomous driving are intensifying. There has always been a view that the lack of lidar is the reason for Tesla's frequent accidents.
A federal investigation into a fatal crash involving a Tesla in Florida in 2016 revealed that the Tesla’s own sensors and cameras failed to recognize vehicles under bright skies. The owner of the white semi-truck did not put his hands on the steering wheel. As a result, the Tesla in autonomous driving mode crashed into the truck at high speed.
After the accident, Richard Wallace, director of the Transportation System Analysis Group at the American Center for Automotive Research, said that although the Model S is not designed to achieve full autonomy, it is still recommended that Tesla consider using lidar to achieve full autonomy. drive.
The root cause of drivers’ abuse of Tesla’s autonomous driving system lies in the public’s over-perception of Tesla’s autonomous driving level. In fact, Tesla is still at the L2 semi-autonomous driving level and still requires the driver to control the steering wheel.
To move from L2 level to higher level autonomous driving (L3 and above), lidar has become the currently recognized best route. Compared with the "pure vision solution", LiDAR has the characteristics of long range measurement, excellent angular resolution, and is less affected by ambient lighting. It does not require deep learning algorithms and can directly obtain the distance and orientation information of objects.
According to Sullivan’s prediction, the global market size is expected to reach US$13.54 billion in 2025. In terms of the domestic market, Sullivan predicts that China's lidar market will reach US$4.31 billion in 2025, with a compound growth rate of 63.1% from 2019 to 2025, of which the automotive field is the main component.
Ni Zhengyang, an analyst at Debon Securities, said that the performance advantages of lidar + price drop + urgent downstream demand are the three factors that have promoted lidar to become a mainstream solution for mass production of automobiles.
However, the industry is still in its early development stage, and both Chinese and foreign manufacturers have strong competitiveness. Foreign manufacturers such as Velodyne, Luminar, and Aeva are taking the lead in mechanical, semi-solid, and FMCW technology paths, while Chinese manufacturers Hesai Technology, Sagitar, Huawei, DJI Livox, etc. also have strong competition with their own high-quality products. strength.