The competition between China and the United States in the autonomous driving industry: A big competition in the industrial chain, who has more leadership?

Hello, everyone. I am your Ying Ge. Today I am going to talk to you about things related to autonomous driving in China and the United States!

First of all, a very objective fact is that in the traditional automobile industry, including design, manufacturing, R&D, etc., there is a relatively large gap between China and Europe and the United States, and it is very difficult to overtake in the short term.

However, the rapid development of the new four modernizations has given us an opportunity to keep pace with European and American companies. Whether it is new energy or intelligence, the technology gap between China and the United States has not widened to a large extent. In fact, we still have an advantage in some areas.

Take autonomous driving for example. Currently, the recognized top players are basically China and the United States. It is not an exaggeration to say that "the development of autonomous driving depends on the development of autonomous driving in China and the United States."

Today, we will focus on the competition between the autonomous driving industries of China and the United States to see who is the leader in autonomous driving!

The industry chain of autonomous driving can be divided into three levels: perception, decision-making and execution. Perception mainly corresponds to cameras and radars (including ultrasonic/millimeter wave/lidar); the decision-making level mainly corresponds to chip algorithms, high-precision maps, V2X communications and other aspects; and the execution layer is the last link, which can also be understood as the whole vehicle application.

Comparison of the autonomous driving industry chain between China and the United States: Perception layer - on-board cameras

Let’s talk about the perception layer first. The general consensus in the industry is that before visual technology achieves a breakthrough, autonomous driving vehicles must use cameras, ultrasonic radars, millimeter-wave radars, and lidars as perception devices.

We analyze them one by one, starting with the car camera. The industry chain of the car camera is as follows

Among them, CMOS (photosensitive element) is the core component of the entire camera. The industry is basically monopolized by American, Japanese and Korean companies. The absolute leader in the industry is the American company On Semi, which has a market share of 46% in the automotive image sensor market, while domestic companies are still in the starting stages in this field.

In the field of lens sets, China's Sunny Optical leads the global market in automotive camera lens sets, with a market share of 39% in 2017.

Its products include front-view cameras, surround-view lenses, rear-view lenses, interior-view lenses, etc. Its products have excellent characteristics such as continuous and stable focusing, good thermal compensation, and high light transmittance, and can be widely used in various autonomous driving scenarios.

In terms of software algorithms for vehicle-mounted cameras, Mobileye, a subsidiary of Inter, is basically the only dominant player, with a market share of around 90%.

Simply put, in the field of automotive cameras, Chinese companies still have a relatively weak presence compared to American companies, and are generally catching up.

Comparison of the autonomous driving industry chain between China and the United States: Perception layer - millimeter wave radar

At present, automotive millimeter-wave radars are mainly concentrated in 24Hz and 77Hz (will converge to 77Hz in the future). From the perspective of the market structure, the top players have basically been monopolized by companies in Europe, the United States, Japan and other countries.

Among them, Delphi's millimeter-wave radar is mainly based on 77GHz, using a more traditional hardware solution, which is relatively expensive, but has good performance.

Domestic companies started late in the field of millimeter-wave radar. Currently, the product system of 24GHz millimeter-wave radar in the domestic market is relatively mature and the supply chain is relatively stable, but 77Hz millimeter-wave radar products are not currently perfect, and core radar chips are restricted by foreign companies.

Currently, the main domestic suppliers of millimeter-wave radars include Beijing Xingyidao (77GHz), Desay SV (24GHz, 77GHz), Shenzhen Anzhijie Technology (24GHz, 77GHz), Hangzhou Zhibo Technology (24GHz, 77GHz) and other companies.

In general, in the field of millimeter-wave radar, the domestic industrial chain is not as complete as that of the United States and other regions, and its radar core chips are still restricted by other countries.

Comparison of the autonomous driving industry chain between China and the United States: Perception layer - LiDAR

As one of the key sensing devices for achieving high-level autonomous driving, lidar is generally considered by the industry to be the most effective environmental perception solution currently.

Speaking of the most famous LiDAR supplier in the industry, it must be Velodyne from the United States. It became famous for providing 64-line LiDAR for Google's driverless cars.

Velodyne currently provides 16-128-line complete sequence lidars with very powerful product performance. Its customers include well-known domestic and foreign companies such as Uber and Baidu.

In addition, Quanergy in the United States is also a major supplier of lidar. It mainly supplies solid-state 8-line lidar, which is much cheaper than Velodyne's mechanical lidar.

There are many domestic manufacturers involved in LiDAR, such as RoboSense and Hesai Technology. RoboSense's most advanced LiDAR has reached 128 lines; Hesai Technology mainly focuses on the field of solid-state radar. Among the 62 high-tech companies in California that have obtained driverless car public road test licenses, more than 40% are already paying customers of Hesai Pandar40.

In addition, the laser radar product released by DJI at CES in January has now been delivered. The product is named Tele-15, which meets automotive-grade standards, has an effective sensing distance of 500 meters, and a service life of 10 years. The price is only 8,999 yuan, which is cheaper than genuine mass-produced products.

Comparison of the autonomous driving industry chain between China and the United States: Decision-making level - autonomous driving chips

If the various sensors mentioned above are compared to our eyes, hands, sense of smell and other sensory organs, then the autonomous driving chip can be compared to our brain, which is responsible for processing various complex calculations.

In this field, China and the United States are basically in a state of chaos. The more well-known American autonomous driving chip manufacturers include Nvidia, Mobileye (mainly used for camera data perception rather than central domain processors), Tesla, etc., while the more well-known domestic chip manufacturers are mainly Horizon and Black Sesame.

Among them, Horizon won the first place in 4 of the 5 challenges in the CVPR 2020 Autonomous Driving Workshop open dataset challenge held by Waymo, becoming famous overnight.

In the field of autonomous driving chips, domestic manufacturers may still have a certain gap with foreign established companies in advanced research and development, but overall the gap is not large. For example, Black Sesame's chip has a computing power close to that of Tesla, but its power consumption is reduced by more than half.

Comparison of the autonomous driving industry chain between China and the United States: Decision-making layer - high-precision map

High-precision maps are considered an indispensable part of autonomous driving. They can obtain prior information for autonomous driving vehicles and improve the safety and comfort of autonomous driving vehicles.

But it is worth noting that high-precision maps involve national geographical security, so relevant licenses are not easy to obtain, and companies engaged in map collection are basically domestic companies.

Well-known Chinese map providers include Baidu, AutoNavi, and NavInfo. Baidu Maps has completed mapping of 300,000 kilometers of expressways and some urban roads, with a relative accuracy of 10-20 centimeters. AutoNavi has currently completed more than 320,000 kilometers of high-precision map collection for autonomous driving, with an accuracy of up to 10 centimeters.

High-precision map suppliers in the United States include Mobileye, Waymo, TomTom, etc. Among them, the high-precision maps developed by Waymo have met the L4 level autonomous driving requirements of many cities such as Phoenix, USA (for its own use only), and TomTom's high-precision map data has achieved full coverage of interstate highways and freeways in the United States.

Comparison of the autonomous driving industry chain between China and the United States: Decision-making level - 5G

5G, which has low latency, high speed and wide connection, is destined to be one of the prerequisites for the full implementation of autonomous driving. In this regard, China is ahead of the United States in all aspects, whether it is related technology, policies or infrastructure.

Friends who know something about the United States must know that the U.S. telecommunications infrastructure is not perfect. Signals are not fully covered in many areas. Outside the city, there are basically mountainous areas. Even the 4G signals in many urban areas are not good, let alone 5G.

For China, which is known as an infrastructure maniac, this is not a problem. By 2025, there will be 430 million 5G connections in my country, and China will basically achieve full coverage of 5G signals. This is undoubtedly major good news for the implementation of autonomous driving.

Comparison of the autonomous driving industry chain between China and the United States: Competition among leading companies

If we count from 2009 when Google launched its driverless car project Project Chauffeur, humans have already gone through the first 10 years on the road of autonomous driving.