Recently, Li Bin, CEO of NIO, talked about his views on lidar at the Science and Technology Day, he said, "To say that lidar is useless must be either stupid or bad, because lidar plays different roles at different stages, and Tesla is not right about everything." ”
In fact, there have always been two solutions in autonomous driving technology, one is the pure vision solution that Tesla insists on, and the other is the lidar solution.
In fact, these two schemes have their own advantages and disadvantages, and the "LiDAR solution" can detect the surrounding information regardless of good weather, and does not require complex calculations, but it does not look more intuitive without a pure visual solution.
Most of our domestic cars use a comprehensive lidar scheme, that is, lidar-based, which also includes millimeter wave (including more advanced 4D millimeter wave radar), cameras and other auxiliary equipment, so it has made some compensation intuitively, and the effect is also very good.
The "pure vision solution" is to imitate the human eye, so it needs to be very powerful and computing power, but the camera is susceptible to the weather, and there is still a large distance between it and the human eye's precision and reaction speed, but the principle is indeed easier for people to understand and accept.
Some people even made an altitude map of intelligent driving technology, and Tesla and NIO are in the first echelon.
Of course, Li Bin also mentioned that Tesla was the first car company to do autonomous driving, because when the Model Y/S was built, foreign lidars cost tens of thousands of dollars a piece, and the lidars at that time were rotary.
Indeed, since the lidar technology has been mastered by us, the price has simply fallen from the sky to the ground, tens of thousands of dollars have dropped to a few thousand yuan, and now the highest and most expensive lidar in our country is only a few thousand yuan, and some are even only one or two thousand yuan.
In Li Bin's view, as long as the information provided by LiDAR is useful, accurate and reliable, it is useful, and whether it is used or not is a business problem for enterprises, and there is no problem with technical routes.
Xiaobian thinks that Li Bin's statement is not unreasonable, at present, domestic car companies are using a large number of lidar, such as Wenjie, Xiaopeng, Weilai and other car companies are using this scheme, which shows that lidar is not as unbearable as Musk said in 2021. Because the domestic product is not a pure lidar route, but a combination equivalent to a lidar + vision solution, which can make up for each other, improve the accuracy at the same time, and greatly enhance the intuitiveness.
Speaking of NIO, it has not burned less money in recent years, and the total investment in R&D reached 13.431 billion yuan last year, becoming one of the most willing car companies among the new power car companies.
In recent years, NIO has built 12 #full-stack self-developed #technology-leading layouts, and has high-quality results in both software and hardware.
NIO has developed core technologies such as Tianshu SkyOS, the first full-vehicle global operating system of an automotive company, the first chip product LiDAR main control chip "Yang Jian", NIO Link, the mass-produced ADAM CCC cross-domain fusion supercomputing cluster, the first and successfully tape-out 5nm automotive-grade intelligent driving chip Shenji NX9031, the global 900V high-voltage architecture, and the SkyRide intelligent chassis system. It can be said that it almost covers the software and hardware technical capabilities of smart electric vehicles, as well as the strong digital operation capabilities that support products, services, and communities.
Let's take their self-developed and mass-produced lidar main control chip "Yang Jian" as an example, the chip is equipped with an 8-core 64-bit processor, as well as a sampling depth of 8 channels and 9bit, and a sampling rate of 1GHz. While efficiently capturing the raw data of the LiDAR sensor, it also reduces the power consumption and 30% latency of the LiDAR by 50%, and the point cloud processing capacity is 8 million per second, which doubles the amount of point cloud data per frame, providing richer data support for accurate mapping, and providing more efficient and accurate data even in more complex or even worse driving scenarios.
In addition, NIO's self-developed 5nm automotive-grade intelligent driving chip Shenji NX9031 is also very powerful, and it can be worth four at one, what does it mean? In other words, the performance of the four flagship intelligent driving chips in the industry can be achieved with one Shenji, which not only reduces the hardware cost, but also brings users the most efficient intelligent driving experience.
According to data, as of the end of June 2024, the total number of patents applied for and authorized by NIO worldwide has reached 9,000, and the amount of R&D investment and the number of patents of NIO have ranked first in the list of new forces. #蔚来技术#
In short, whether it is a lidar solution or a pure vision solution, as long as it is the same, it is a good technical route.