The Battle of Self-driving Cars between China and the United States

The battle of self-driving cars between China and the United States has been simmering for ten years. However, the scales tilted unexpectedly in a short month in 2022.

On June 28, 2022, in the dead of night, several streets in San Francisco, California, suddenly became chaotic. The empty streets were clogged with taxis from the self-driving company Cruise, causing a traffic jam. Contrary to expectations, this did not herald a breakthrough in Robotaxi’s business. According to an in-depth interview with the well-known US technology media outlet Wired, Cruise’s fleet suddenly lost contact with the backend server, leading to collective crashes. The chaos on the street and in the control rooms is a microcosm of the US Robotaxi flagship.

Meanwhile, on the other side of the Pacific, footsteps from China are moving from chase to surpass. One month later, the first batch of unmanned commercial policies went live in China. Fully unmanned Robotaxi self-driving services with paid travel services officially launched in Wuhan and Chongqing. The continuous upgrading and opening of regulatory policies have provided a wider growth space for domestic self-driving enterprises, and an outstanding opportunity for China to overtake the United States on the winding road of autonomous driving.

Baidu, who has long been regarded as the flagship of Chinese Robotaxi, has become the only operating company in the two cities.

The Self-Driving Chronology of Two Cities

The grass isn’t always greener on the other side.

A few years ago, in Chongqing and San Francisco – two cities separated by the Pacific Ocean – the remote battle of self-driving technology was in full swing.

In March 2016, General Motors purchased Cruise, headquartered in San Francisco, for $1 billion. This was the largest acquisition case in the Silicon Valley automotive industry at the time, which fully demonstrated General Motors’ determination to develop autonomous driving. Nine months later, Waymo, a self-driving project under Google, became independent. Since then, the two world-leading self-driving companies have entered a new era of development. Some say that 2016 was the year the world accelerated the development of self-driving technology.

At that time, Chongqing had just begun to build China’s first unmanned driving test demonstration area.# Cruise’s Third-Generation Self-Driving Car

In October 2017, Cruise began road testing their third-generation self-driving car in the main urban area of San Francisco. It is worth mentioning that this was only two months after their second-generation autonomous vehicle began internal operations and 14 months after the release of their first-generation self-driving car.

Just a few months prior, Baidu had just released their own autonomous driving platform, Apollo.

In 2018, Cruise began testing self-driving taxis in San Francisco.

In April of the same year, Chongqing issued the first batch of eight autonomous vehicle test licenses, with Baidu being granted one.

From around this time, the slope of the autonomous driving development curve began to quietly exchange between the two cities.

In 2019, Cruise faced challenges in regulation and technology, with the company’s plans to deploy self-driving taxis on a large scale in San Francisco falling through for the year.

In contrast, 2019 was a great year for Baidu as their autonomous driving technology was successfully launched in 23 cities throughout the country, with a total testing mileage of over three million kilometers. More importantly, Baidu Apollo has maintained an outstanding safety record of zero accidents.

In the following years, both China and the United States roughly maintained the above development momentum until the scene at the beginning of the article happened this year.

All of this is worth reflecting on.

The Sound of the Eve of Autonomous Driving

“Advanced science and technology are the primary productive forces.” This is a sentence we have learned since middle school, and in the technological panorama of the 21st century, AI artificial intelligence is definitely one of the most important puzzle pieces.In fact, AI has already permeated many aspects of our daily lives, from improving the quality of photos taken by smartphones to assisting in epidemic prevention and promoting the reform of Industry 4.0, and the importance of AI technology is self-evident.

The functions of environment perception, prediction, planning, and control derived from autonomous driving technology all require the power of AI. For AI technology, the development of autonomous driving is of great significance.

As early as two years ago, China announced the “Roadmap for Intelligent Vehicles Development 2.0” at the 2020 World Intelligent Connected Vehicles Conference, laying out top-level planning for the development of intelligent driving in the next 15 years: “By 2025, the penetration rate of L2 and L3 intelligent driving in new vehicles should reach 50%, and by 2030, this number should exceed 70%, and the penetration rate of L4 should also reach 20%.”

However, in the short term, the high cost of core components such as computing chips and lidar for advanced autonomous driving will greatly limit the development of the industry.

Regarding this issue, the world’s leading data provider, IHS Markit, has made predictions that if national-level incentive measures achieve positive results in the market, and economies of scale rapidly emerge in technical costs, the penetration rate of autonomous driving will rapidly increase.

For privately-owned vehicles gradually upgrading from L2 to L4, commercial Robotaxis at L4 level are an excellent supplement to the core component costs.

Perhaps this is one of the main reasons for the policy’s openness, and Baidu’s commercial operation of truly driverless Robotaxi will also be the herald of the arrival of large-scale autonomous driving.

Why Baidu?

When it comes to autonomous driving technology in China, it is impossible to avoid talking about Baidu. Since 2013, Baidu has been laying out autonomous driving and has developed to the point where we can confidently say that China’s Baidu has surpassed the United States in terms of products, technology, cost, and operational scale in the field of autonomous driving. Baidu has also become the largest global provider of autonomous driving travel services at present.

Let’s forget about the distant future and focus on what’s freshest.

When other automakers are still struggling between 0, 1, 2, or 3 LiDARs, half a month ago, Baidu launched the sixth-generation autonomous vehicle, Apollo RT6, at the “Baidu World Conference 2022,” which instantly increased the number of LiDARs on a single vehicle to eight.

Obviously, for the L4-anchored Apollo RT6 to face Chinese road conditions, it needs to achieve precise 360° environmental perception. The complexity of Chinese road conditions is far beyond that of American large rural areas. According to statistics, the number of participants and complexity of Chinese roads are 15 times that of the United States. Therefore, compared with their American counterparts, Baidu’s use of eight LiDARs on the RT6 demonstrates their technological leadership in this field.

What excites us more than the number of LiDARs is the cost. Don’t forget the “mass production” prefix adjective in RT6. Facing mass production, Baidu claims that the RT6’s overall vehicle cost will decrease to 250,000 yuan, which is a 48% decline compared to the cost of Apollo Moon, the previous generation autonomous vehicles, which was 480,000 yuan. This is almost a halving of costs.

Compared with foreign autonomous driving companies, such as Cruise and Waymo, whose vehicles cost millions of yuan, Baidu RT6 achieves better results at one-fourth, or even lower, of the cost, which constitutes devastating blows in terms of cost-effectiveness.

Behind the efficiency of cost reduction is the wave caused by technological innovation.

Baidu’s Apollo, the world’s first open autonomous driving platform, launched in 2017, has now become the world’s most active autonomous driving platform, bringing together over 210 global ecologic partners and 80,000 developers. Baidu has publicly disclosed more than 3,700 patents in autonomous driving, making them the leader of this field in China.

Up to now, Baidu not only owns 672 automatic driving test licenses in China, but also has autonomous passenger vehicle test permits in California, USA. The size of Apollo’s test fleet has reached 500 vehicles, and the total actual mileage on city roads worldwide has exceeded 32 million kilometers.

The strong technological foundation translates into commercial success, which is reflected in the rapid development of Baidu’s autonomous driving transportation service platform, “Apollo Go”.

“Apollo Go” has already been launched in more than 10 cities in China, such as Beijing, Shanghai, Guangzhou and Shenzhen, and accumulated more than 1 million orders in just the past year. Meanwhile, the annual order volume of Waymo, the largest Robotaxi service provider in the United States, is only about 160,000 orders. In terms of order volume, Baidu Apollo has formed an overwhelming advantage over American Robotaxi companies.

With the lower-cost Apollo RT6 becoming available, “Apollo Go” is expected to accelerate its expansion again. Baidu expects to operate in 65 cities by 2025 and cover 100 cities by 2030.

We believe that soon, unmanned taxis will be available to more ordinary people.

Where is Robotaxi Heading?

Until today, our roads are still made up of independent individual vehicles that are put together from the bottom up. A city with dozens or even hundreds of thousands of drivers has hundreds or even millions of different driving behaviors. When these behaviors interfere, intersect, or even deadlock with each other, congestion and accidents occur as a result.

But imagine if all the vehicles on our city roads become Robotaxis, using V2X communication technology to communicate with the external environment, and AI big data allocation to optimize traffic flow, the entire city traffic can be switched to a top-down integrated operation mode. In this mode, users only need to express their travel needs, and the best solution will be coordinated and calculated by the system.

This scenario has been rehearsed in Baidu’s mind countless times.

Robin Li said, “Intelligent transportation based on vehicle-road collaboration can improve traffic efficiency by 15%-30%, and drive an absolute increase of 2.4%-4.8% in GDP every year. With the deepening of intelligent transportation practice, first-tier cities in China will no longer need to restrict car purchases and traffic as well, and intelligent transportation systems are also expected to reduce traffic accidents by 90%.“

Robin Li implanted the advantages of Robotaxi into our minds with a set of data.

Efficiency, economy, and safety are probably the smooth world that every person who has experienced heavy traffic yearns for.

And this beautiful vision may start today, August 8, from Baidu, from Chongqing, Wuhan, and China.

This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email bd@42how.com.