Changes in travel in 2020 (I): The “leap into the abyss” moment for autonomous driving

In 2020, with the maturity of new digital technologies represented by 5G, AI, and cloud computing, coupled with the impact of the COVID-19 pandemic, almost all traditional industries and offline scenarios have been forced to accelerate their digital and intelligent transformation and upgrading, especially in the field of transportation.

In the field of transportation, the three major industries of autonomous driving, new energy vehicles and intelligent transportation have achieved a series of new technology and scenario innovations in 2020. Fully autonomous driving has been put into commercial use, new energy vehicles have seen a rise against the trend, and intelligent transportation has begun to promote the vehicle-road collaboration scenario of 5G Internet of Vehicles.

Standing at the beginning of 2021, we hope to make a panoramic review and summary to mark the changes in travel scenarios such as people, cars, roads, and the Internet in the past year.

Let’s talk about 2019. The progress of autonomous driving technology in this year was not satisfactory. Startups experienced a reshuffle, and traditional car companies were unable to support themselves. They began to band together and lowered their expectations for self-developed autonomous driving technology. It can be said that the autonomous driving industry has ushered in its first wave of troughs since its outbreak in 2016.

In 2020, the whole world was hit by the black swan of the COVID-19 pandemic. The high risk of human-to-human transmission and the pain of home isolation have made us look forward to the arrival of self-driving cars even more. However, ordinary people can only see self-driving cars in the news. We are still far from the ideal of summoning a driverless taxi (Robotaxi) like driving a shared bicycle.

How far is the dream?

A few years ago, many self-driving startups and automakers entering the self-driving market had given 2020 many overly important missions. For example, fully self-driving cars were running all over the streets, and many manufacturers' L3 self-driving cars were also launched on a large scale.

2020 still disappointed everyone. However, reality is never as good as expected, but it is not too different.

For most countries, L3 autonomous driving will not be legally available on the road in 2020, and many car companies' so-called "first L3" can only be upgraded to L2+. However, the more optimistic trend is that L4 advanced autonomous driving technology has been implemented in the Robotaxi travel projects of many autonomous driving companies. People in some cities in China and the United States can use their mobile phones to call a self-driving taxi. Moreover, the most critical progress in 2020 is that the safety officers have been removed from the commercial vehicles of these Robotaxi projects. Passengers can finally experience sitting in the back seat, looking at the steering wheel with a little excitement and "freely" playing, and then take you to your destination.

So, what kind of time node should 2020 be considered for autonomous driving?

在《周易·易经·》中，乾卦爻辞的三、四爻辞这样说：“九三，君子终日乾乾，夕惕若，厉无咎；九四，或跃在渊，无咎。”我们知道，《易经》的乾卦描述了事物在不同阶段的变化特征。九三是一个处处危机、小心谨慎的“惕厉”时刻，而九四则是事物开始摆脱困境，可上刻下的“跃渊”时刻。

In the context of the development of autonomous driving in recent years, 2019 can be regarded as a "vigilant" moment for the autonomous driving industry, while 2020 can be regarded as a "leap over the abyss" moment to get out of the trough and begin to usher in a bright future.

French philosopher Leibniz once said that "there are no leaps in nature" and all changes are in a continuous chain.

This also applies to the development of autonomous driving technology. It is difficult to say that there is a critical "singularity" moment between "vigilance" and "leap into the abyss". For example, Waymo quietly started testing without a safety officer in 2019. The difference in 2020 is that the autonomous driving Robotaxi mode without a safety officer is actually on the road and officially appears as a commercial product.

2020 was also difficult for the autonomous driving industry, but fortunately, under the epidemic, it also ushered in an upward growth situation. This is the keyword we use to describe this year: "or leap into the abyss."

Looking back at 2020, we are concerned about the three aspects of the autonomous driving industry: the industrial climate, technological level, and commercial boiling point. Let’s get to the point.

Industry climate: Relying on autonomous driving to sell cars VS relying on autonomous driving to carry passengers

We know that there are two tracks for autonomous driving. One is gradual, starting from L2 and gradually moving towards L3 and L4 advanced autonomous driving. Tesla and other car manufacturers are representatives of this model. Selling cars is their main business, and autonomous driving is currently an added value. The other is radical, starting directly from L4 and achieving fully unmanned driving. Waymo, Baidu and other autonomous driving companies are representatives of this model. Robotaxi travel is their main business, and selling technology to vehicle manufacturers may become their main business.

The year 2020 will be the intersection of these two tracks. Players on the low track are seeking to move towards L4 advanced autonomous driving, while players on the high track are reducing their dimensionality to the low track. When the main business is difficult to generate revenue, they are seeking more cooperation to realize technology monetization.

What’s interesting is the awkward transition level of L3. In March 2020, Audi, which was the first to launch L3 cars, officially abandoned the L3 autonomous driving project and switched to the research and development of L2 and L4 autonomous driving projects.

L3 is technically possible, but once it is put into practice, its feasibility is even worse than L4. This is because L3 autonomous driving has a difficult-to-solve "takeover paradox", that is, when the L3 autonomous driving system cannot work properly, the driver who is distracted, resting or even sleeping, how to take over and when to take over the vehicle.

Either keep paying attention throughout the process like L2 level, or achieve fully autonomous driving at L4 and L5 levels. It is impossible to set this "Schrödinger's cat" state for the driver or passengers, because once an accident occurs, it is impossible to judge the responsibility of the driver and the autonomous driving platform.

Therefore, in the autonomous driving industry, we will see that cars that are close to or have L3 capabilities will still be put on the market with L2 combined driving assistance scenarios, or directly put the car on the road without L3 level, which means that autonomous driving companies bear 100% responsibility for the driving safety of the entire vehicle.

The first idea can be seen from Tesla's FSD Beta update. By upgrading the FSD kit of Tesla with L2 autonomous driving capabilities, it is possible to achieve automatic assisted lane change, automatic assisted driving navigation on the highway, identify traffic lights and respond to parking signs. In the parking lot, Tesla can automatically park into place, and the owner can also perform smart summons. Is it good? Recently, a Tesla owner has personally tested it and completed a fully automatic driving from Los Angeles to Silicon Valley with a Model 3, taking over for more than 1 minute when stopping to charge.

It has evolved from L2 to L4 advanced autonomous driving, but before the technology is mature and regulations are finalized, it is still launched into the market at L2 level. This is exactly the goal that the new domestic car manufacturers are striving to catch up with.

The second idea is what we are seeing from Waymo, Baidu, and other companies planning driverless travel.

Last October, Waymo CEO John Krafcik officially announced that Waymo One users would be able to use fully driverless taxi services. No longer equipped with a safety officer, it means that even if there is an emergency on the road, the autonomous driving system is responsible for the safety of the vehicle and passengers. In Guangzhou, Changsha, Beijing, Cangzhou, Shanghai and other places in China, Robotaxi travel by many autonomous driving companies has been conditionally open to the public. At the same time, they have cooperated with many traditional vehicle manufacturers to reduce the dimensionality of advanced autonomous driving technology to L2 assisted driving models to accelerate the commercialization of technology.

As competition in these two fields intensifies, capital and policies have also begun to provide strong support.

Let's look at some major capital moves. In May last year, Waymo completed its first external financing of $3 billion, setting a new global record for a single round of financing for an autonomous driving company. In December, Aurora acquired Uber ATG with equity, becoming the third largest unicorn in the field of autonomous driving. The sale of Uber's autonomous driving business can only be seen as its own failure, and the integration of autonomous driving Robotaxi and shared online car-hailing business is still the most commercially valuable track for autonomous driving.

In China, from the beginning of the year to the end of the year, autonomous driving companies such as Pony.ai, AutoX, Didi, and WeRide have also received hundreds of millions of dollars in financing. Consortium capital, technology capital, and OEMs are still the main investors in these companies, which also indirectly confirms the industry's optimistic expectations for the future of the autonomous driving industry.

Let's talk about the policy support from China and the United States. In 2020, the epidemic did not hinder the development of industrial policies. The United States released "Autonomous Driving Vehicle 4.0" in January last year, clarifying the US government's promotion of autonomous driving technology innovation. In February last year, my country released the "Intelligent Vehicle Innovation and Development Strategy" and "New Energy Vehicle Industry Development Plan (2021-2035)", which established the development goals of autonomous driving by 2035 and promoted the large-scale application of L3 and L4 autonomous driving vehicles.