Starting today, the driver's seat is no longer a human seat, and starting from Beijing, even the main driver safety officer responsible for monitoring and emergency takeover has been removed.
April 28, 2022, Beijing, Yizhuang.
A day of wind and sunshine, a sparse and ordinary day, but from now on it will surely be a day that will be recorded in the annals of history.
Robotaxi, the kind of driver who has no one, is officially allowed to carry passengers on the road.
The "no driver behind the wheel" travel experience, once a sci-fi movie plot, has become a reality in China.
Baidu once again took the lead in winning the license license, and Beijing citizens can call driverless Robotaxi through its Radish Run App.
But the issuance of unmanned Robotaxi manned permits may be of more significance than policy and progress itself.
Behind it is the overall speed and process of China's automatic driving, which has achieved a landmark overtaking on the other side of the ocean, from going hand in hand to the leading stage.
What's different about riding an unmanned Robotaxi?
Beijing passed the test with unmanned vehicles on the road passenger trial operation with unmanned manned demonstration application license.
That's right, you'll have to pass a demanding technical exam before you can officially hit the road.
And earlier, the relevant real road test of unmanned manned man has been paving the way, after Beijing issued the implementation rules for unmanned road test, allowing eligible players to take off the safety officer.
Now, after feedback and guarantee from real-way test feedback and security management, network, data security, and risks, unmanned manned scenarios have become a reality.
We tried it at the first time, and there are three points of experience that we can share:
First, the taxi process.
Second, the experience of the itinerary on the road.
Third, the successful resolution of special scenes.
Let's start with the taxi process.
It's the same as experiencing Robotaxi before, or it's no different than calling a ride-hailing car.
Open Baidu's Radish Run App and select the destination site you want to go to.
After waiting at the starting point, there will be a Robotaxi who has no one in charge to pick him up.
After the car stops, scan the code to open the car door, get on the back seat, fasten the seat belt and click "Start Trip" on the big screen.
There are large Pad screens in front of the rear seats, which not only display travel information and real-time road condition perception information, but also choose to play music, watch videos, and adjust the temperature of the air conditioning in the car.
However, according to the current policy requirements, passengers can only ride in the back seat at present, because the front passenger needs to be equipped with a safety officer.
However, the difference between what the co-pilot safety officer can do and the previous main driver safety officer has also been huge.
The main driver can take over the steering wheel, slam on the brakes or operate like a normal driver, while the main ability of the co-pilot safety officer is almost the brakes – somewhat similar to the driving school instructor.
Along the way in our experience, the co-pilot safety officer "did nothing".
So what was the experience on the road?
First of all, the overall body feel is smooth.
Although Robotaxi has been rid of human driver dependence before, but the real "emptying" of the driver's seat will look and feel very different.
Empty position, automatic steering wheel... Depending on the speed and braking of the road conditions, the experience is quite different from when there is a safety officer.
But the overall opening is very smooth, if you sit in the back seat of the main driver, because the line of sight will be blocked by the Pad, it is not likely that the front row is empty.
Several major projects that generally examine the level of autonomous driving technology, such as traffic light recognition, unprotected left turn, pedestrian and other traffic subject detection, have not been able to pose challenges for AI drivers.
Secondly, the overall experience will also feel secure, that is, a sense of security.
On the one hand, the rules and safety from AI drivers are supreme, such as slightly decelerating when crossing intersections, and there will be avoidance strategies for dense traffic lanes, large cars and pressed lines.
On the other hand, the UI interaction provided by the large screen of the Pad synchronizes the "global information" to the passengers in real time, where they are, where the itinerary is going, and then turn left and right... At a glance, it's clear.
During the trip, I was once surprised by a two-wheeled electric car speeding by, but Baidu's AI driver seems to have already seen it strangely, and the waves are not alarming...
After the trip, the safety officer on the co-driver was asked to introduce in time, in a similar and special scene, the AI driver will even honk his own horn...
Finally, there were special road conditions along the way -
traffic accident.
Two large cars rubbed in the right lane, not only occupying the rightmost lane, but also pressing the lane change line... Just added to the blockage of the unmanned Robotaxi who needed to turn right.
Unexpectedly, Robotaxi had a slight "hesitation", parked "observed" for a few seconds, and turned out to be... Circled around to complete the right turn.
Amazing……
In short, the entire unmanned Robotaxi passenger experience is no different from the main driver with a safety officer and a human driver.
Even as an electric vehicle, the rear seat is gentle and comfortable, higher than many human online ride-hailing drivers who drive electric vehicles.
It can be expected that after more and more ordinary citizens ride the experience, the freshness of the main driver will soon dissipate.
As a tool for travel transportation, it is a bit unpretentious.
But even so, behind this unmanned Robotaxi implementation, it is not easy.
What's so hard about the driver going to people?
At least three conditions are required, one of which is indispensable.
First, technical capabilities.
Second, industrial maturity and market acceptance.
Third, policies and regulations.
Moreover, as the most eye-catching part of the intelligent transportation revolution, Robotaxi's unmanned manned landing is actually an all-round multi-dimensional comprehensive strength competition and racing.
The first is the most core technical capabilities.
Specifically, the technical capabilities of completely unmanned driving must not only drive well, stably and safely under the most complex urban open road conditions, but also have the ability to make decisions and process various long-tail scenarios.
Whether there is a person in the driver's seat or no one is fundamentally different for the technical system.
Previously, Li Zhenyu, senior vice president of Baidu Group and general manager of the intelligent driving business group, summarized this technical ability with AI old drivers.
He said that the autonomous driving system can achieve continuous automatic driving in different road conditions, scenarios and areas like the old human driver.
This kind of technical ability, to the operation and landing scenarios, but also with a more macro overall perspective.
Operational and vehicle engineering capabilities, including 5G cloud driving and front-loading mass production vehicles, are also indispensable in the actual landing.
In this unmanned Robotaxi experience, the successful bypass of the front accident vehicle we mentioned earlier, and then the Baidu staff gave feedback - the background system has early warnings and prompts, so the cloud driver in the rear has intervened.
The second is industrial maturity and market acceptance.
Before Robotaxi entered the scale and commercialization, industry and market-related factors did not receive the same attention as technology.
But after entering the landing and commercialization, is the industry behind Robotaxi mature enough?
For example, whether the new energy automobile industry is perfect, and whether key sensors such as lidar have the ability to mass-produce on a large scale, all need to become issues that must be considered in the development of Robotaxi.
Similarly, there is market acceptance.
The main thing is whether the user trusts and is willing to pay for it, because the demand side of the relationship directly leads to the huge R & D investment in the early stage of automatic driving, and whether it can really achieve returns in the expected broad market scale.
Finally, the implementation of unmanned Robotaxi manned, the most important thing is the relevant policies and regulatory permits.
This is also why the issuance of licenses in Beijing is so worthy of attention.
In the early development of autonomous driving, California, which has always been a relaxed state, has led the rest of the world, laying the groundwork for the barbaric growth of many players.
However, after the inflection point of automatic driving and commercialization, China not only attaches importance to the development and landing of automatic driving at the national level, but also has a high design and drive at the top level.
In different parts of the country, there is also a development trend of first-mover and scramble, including a number of first-tier cities such as Guangzhou, Shanghai, and Changsha, which have given policy guidelines and landing guidelines for automatic driving.
The unmanned Robotaxi passenger license in Beijing reflects the important role of policy licensing for the landing and promotion of automatic driving.
For example, before the unmanned Robotaxi manned test, Baidu Apollo has actually proved the maturity of unmanned technology step by step under policy guidelines and permission:
First, unmanned testing and manned trial operation were realized in the semi-closed Shougang Park;
Subsequently, unmanned tests were carried out on the open sections of Yizhuang that met the requirements;
Finally, after continuous safety testing, technical assessment, and safety assurance at the operational level, it was approved for trial operation of passengers.
So overall, the unmanned Robotaxi landing, from technology, market to policy, elements are indispensable.
And it is also the racing nature of this comprehensive strength that also makes the global jianghu pattern of autonomous driving show subtle changes.
China's autonomous driving landing speed is leading the way.
Baidu Apollo, a Chinese self-driving goose player, is also surpassing the American self-driving goose player Google Waymo.
China's historic overtaking of autonomous driving
Unmanned Robotaxi manned opening, Beijing is not the first, but once you enter Beijing time, you can see the micro-knowledge.
It is recognized in the industry that Robotaxi is the crown jewel of autonomous driving, and the unmanned Robotaxi manned vehicle is the prelude to the complete demonstration of Robotaxi's ultimate business model.
Therefore, the unmanned Robotaxi manned node is naturally also a key node.
Being able to land in Beijing is a key indicator among key nodes.
Another industry recognition is that considering the level of AI veteran drivers, it is also necessary to know the world and adapt to local conditions.
Landing in a densely populated area with many transportation elements is much more technically challenging.
There have been comparative analyses before, the participants and complexity of China's roads are 15 times that of the United States, and even after the California DMV report lost its reference power, Beijing's annual autonomous driving test report has become the most "gold-containing" representative because of its meticulous granularity.
Therefore, the unmanned Robotaxi manned landing in Beijing not only means that the technical conditions are mature, but also represents that in the wave of Robotaxi landing in the air race, China has begun to take the lead.
In addition, this is only the potential energy and initial speed of China's autonomous driving from the technical level.
As mentioned above, the race for autonomous driving is becoming a comprehensive power competition that includes industry, market and policy.
The industry, market and policy are the predictable guarantees for China's automatic driving acceleration.
In terms of industry, China is not only the largest market for new energy vehicles, but also the country with the most active related innovations, with a complete new industrial chain and obvious advantages in factors such as cost, manufacturing and data iteration.
As well as the key sensor lidar needed by Robotaxi, Chinese players have shown dominance in terms of cost, supply and mass production.
Of course, China's demographic dividend and acceptance of new things have once again shown unparalleled force.
In 2021, J.D. Power, a well-known foreign analyst agency, conducted a consumer confidence index survey on automatic driving, and the results showed that the confidence and expectation index of Chinese consumers in autonomous driving technology reached 50 points.
In contrast, the confidence index of American consumers in autonomous driving technology has reached only 36.
Another McKinsey study showed similar results.
When asked about the premium they are willing to pay for self-driving cars, Chinese consumers are willing to pay a premium of up to $4,600 (30,000 yuan), compared to $3,800 in the United States.
Therefore, China's existing conditions in terms of industry and market have prepared everything for the unmanned Robotaxi manned to open a larger scale landing.
The opening of scale means that Robotaxi not only realizes the closed loop from technical model to business model, but also means that larger-scale data iteration and system model transition are possible.
So a glimpse of the whole leopard, unmanned Robotaxi manned permission, but also the beginning of China's overall realization of autonomous driving to lead - even if only half a body, even if it is only the first step.
And it is easy to ignore that China still has the strongest support at the policy and infrastructure level.
Like China's space program, China's development and landing of autonomous driving is designed from the highest level and has a global perspective to ensure that it can be achieved.
In the past six years, China has issued more than 10 national policies to ensure the all-round development of autonomous driving from multiple angles such as digital transportation, infrastructure construction, intelligent networked vehicle development, intelligent networking standards, and road testing.
More than 40 provinces and cities, including Beijing, Shanghai, Guangzhou, Shenzhen, Chongqing, Henan Province, Changsha and other provinces and cities, have also introduced more detailed management measures to achieve road testing, manned testing, demonstration operation to unmanned and other stages of division and coverage, to ensure that the definition domain is clear, but also to ensure that the final results are derived according to quality and quantity.
Moreover, in terms of 5G cloud driving and C-V2X vehicle-to-road collaboration, China has also led other regions and become the proposer and promoter of global standards.
Interestingly, players who started out similarly now show a different look.
For example, the respective head geese representatives of China and the United States: Baidu Apollo and Google Waymo.
From the very beginning of the development of autonomous driving, everyone is aiming at Robotaxi, and the core is bicycle intelligent technology.
But now, when I look at it again, I have taken a different path. Waymo is also in bicycle intelligence, Baidu Apollo has realized bicycle intelligence, vehicle-road collaboration, 5G cloud driving, mass production intelligent driving... Including the comprehensive results of intelligent transportation.
Even if it is still around the goal of Robotaxi, Waymo's first-mover lead is obviously unable to withstand the impact of Baidu Apollo.
For example, in the landing of Robotaxi, the three dimensions of technology, product and operation will be focused on.
In terms of technology, the most intuitive measures are mileage, landing richness and road complexity.
In its most recent disclosure (August 2021), Waymo said that the size of the fleet reached 615 vehicles, and the actual road mileage exceeded 20 million miles, equivalent to 32 million kilometers.
Compared with Baidu Apollo's 500 vehicles and 27 million kilometers of real road data, there is still a quantitative advantage.
However, in terms of landing richness, Waymo is concentrated in Phoenix, followed by San Francisco, which began trial operation this year, and no other landing progress was announced.
Baidu Apollo is different, testing in nearly 30 cities, including California, and Robotaxi has opened normal operations in Beijing, Cangzhou, Changsha and Guangzhou.
And the difference in the complexity of road conditions between China and the United States is because of the population density and the richness of traffic participation factors, not in terms of volume, and there have been reports of data before, saying that even a large city like San Francisco may not be able to compare with Beijing and Guangzhou in complexity, let alone the suburbs of Phoenix.
Therefore, the comparison at the technical level, the conclusion is already clear.
At the product level, the focus is on the cost of bicycles - the key to large-scale landing.
Waymo's latest generation of unmanned vehicles, using Jaguar's luxury electric i-pace, plus the entire multi-sensor solution system, cost at least more than one million yuan - this is also the result of Waymo's self-developed lidar and other key sensors.
Baidu's latest generation of unmanned vehicle Apollo Moon has disclosed a cost price of 480,000 yuan, setting a new low in the industry.
At the operational landing level, the most intuitive robotaxi order number, a third party has estimated that Waymo's quarterly number is between 202 and 40,300 orders.
Baidu publicly disclosed in its earnings report last quarter that its radish fast running platform had 213,000 manned orders - at that time, it was also clear that the data was already the first in the world.
As a result, both the technology, product and operational dimensions of Robotaxi are revealing a new pattern trend:
Baidu Apollo has achieved a latecomer to Google Waymo, and China's first goose has begun to lead the world.
And it is foreseeable that with the acceleration of China's unmanned Robotaxi manned policy and more infrastructure in place, the feedback and iteration brought by technology, scenarios and data will further accelerate the above pattern trend.
So today, if you look back at the road ahead, you will also sigh at the fate of Baidu Apollo and Google Waymo, the two major self-driving goose players in China and the United States.
The current results may be the reason for the company's own efforts, but at the same time, it is also the result of the times, the country and various comprehensive forces.
The development of global autonomous driving has come to China's time.