Why is the automatic parking service very popular when the automatic driving encounters a bottleneck? Jiazi Guangnian

Editor | Miss A

Design | Yifan

In the video, a woman gets off near the office building in an Audi A8L. She gently pressed a key in her mobile phone, and the car automatically came to the basement and entered the parking space without driving.

When the lady leaves the office building, she only needs to press a button at the boarding point and the car will drive back from the garage by herself.

Open the door, get on the bus and leave, clean and tidy, eliminating the time-consuming and laborious parking and picking up the car.

This is the scene of "parking service" (the concept of "remote parking" was used in Audi's exhibition at that time).

Functionally, it is an upgrade of the parking assistance system and ADAS (Advanced Automatic Driving Assistance System). Technically, it belongs to low-speed L4 automatic driving.

But in the past five years, the concept of AVP, which is not novel, has not yet become a reality. The upsurge of autonomous driving in recent two years is more manifested in another subdivision track: L4-class autonomous driving of urban road passenger cars.

This is the most consistent with the impression of ordinary people on autonomous driving-the car can automatically carry passengers through the streets and lanes without the driver's operation, and complete all tasks that human drivers should complete, such as overtaking, reversing, avoiding pedestrians, etc.

Its final business model is robot taxi, which belongs to one of TaaS (Transportation as a Service) models. Founded in 2009, Waymo (affiliated to Google's parent company Alphabet) is a pioneer in this field.

Since 20 16, the subdivision of this track has become more and more hot: GM acquired Cruise, an autonomous driving company, and its current valuation has reached 65.438+0.46 billion US dollars; Drive.ai, Pony.ai, Rodarstar.ai, Wenyuan Zhixing and other startups aiming at high-speed L4/L5 autonomous driving have been established one after another; Baidu's Apollo autopilot platform, Uber and Didi's autopilot plans have been launched one after another.

But this "rapid development" seems to be mainly manifested in financing and valuation. After the upsurge of the previous two years, the L4 automatic driving of 20 18 urban road passenger cars entered the bottleneck for one year.

2018165438+10, John Crafts, CEO of Waymo, admitted in public that the popularization of autonomous driving technology will take decades. A few days later, Apple co-founder Steve Wozniak also said: Self-driving cars will not be realized in the near future.

There are only two words that trap these world-class smart minds: mass production and commercialization.

However, AVP is likely to take the lead in breaking the deadlock of mass production and commercialization of high-level self-driving passenger cars in urban space.

Now, the market is only one step away from the mass production of AVP.

It is possible that Mu Zong Science and Technology, a Shanghai company, cut into the field of automatic driving from ADAS assisted driving system.

20 18 12.26, Mu Zong science and technology announced that it has received a letter from China faw, and will deploy low-speed L4-class self-driving AVP products on the mass-produced models of China faw hongqi brand in 2020. Generally speaking, it became the first company in China to win the mass production project of AVP automobile factory.

Let's generally accept orders from the factory. In June last year, they launched an upgraded version of AVP 1.0 for the low-speed scene in the parking lot. The exploration in this field began from 2065438+September 2006.

At that time, in the field of domestic autonomous driving, AVP was not the focus compared with the direction of unmanned vehicles with stronger black technology.

In July of 20 17, Daimler and Bosch announced the concept of "automatic parking service" in the Mercedes-Benz Museum in Stuttgart (their cooperation in this field began on 20 15), and the industry had a more unified view on this scene.

20 18, the AVP game is in full swing.

Looking at the above progress, most of them are the release of concepts, Demo, technical solutions and strategic cooperation.

This illustrates two problems:

First of all, all the players see the prospect of AVP.

Tang Rui, founder and CEO of Mu Zong Technology, also told Jiazi Guangnian: "I think AVP is the most strategic highland in the China market (in the field of autonomous driving)."

Dr Hans Georg Engel, head of Mercedes-Benz China R&D Passenger Car Center, also said last year that AVP is an important milestone on the road of autonomous driving.

Second, there are not many players who have accumulated in this field for a long time.

In the above progress, the research and development of some projects only started last year, and many progress only stays in the stage of cooperative release, which has not yet been tested and mass production.

In this context, the AVP mass production order obtained by FAW Hongqi car in June+February, 5438 is of great market development significance-AVP system is likely to be the first to log in to mainstream passenger cars.

At the moment when many players gathered to March into AVP, Tang Rui, founder and CEO of Mu Zong Technology, reviewed his thinking process to Jiazi Lightyear: Why did you see AVP in the first half of 20 17?

We can look at this problem from the perspective of business logic first.

When considering the opportunity size and commercialization speed of each landing scene of autonomous driving, we mainly look at four dimensions: closed scene/open scene, fixed route/free route, low speed/high speed, and passengers/no passengers on board.

In addition to technical differences, in the above four dimensions, the closed scene/open scene also involves the right of way; Whether there are passengers or not, it is difficult to legislate.

There are two main business models for autonomous driving:

The first one is the TaaS mentioned above. In this mode, autonomous driving is the core technology for establishing operational services.

The second is the traditional business model of the automobile industry, that is, adding automatic driving function to cars and then selling them to consumers. Tang Rui summarized it as VaaP (vehicle as a product). In this model, the company that provides the autonomous driving system acts as the supplier of the automobile factory.

VaaP is now, TaaS is the future.

According to this framework, the most complicated and commercialized situations are "open roads, free routes, high-speed roads and people in cars", which can be directly applied to urban travel-companies such as Roadstar.ai, Pony.ai, Wen Yuan Zhixing and Waymo, the world's top self-driving company, are all challenging this "laurel scene". Robo-Taxi, its corresponding business model, will completely change the way of human travel, and then change the way of life and communication, giving birth to a series of new formats.

This is also the reason why such companies attract a large number of venture capitalists, and the big opportunity should not be missed. But the commercialization time of this scene seems far away at present.

The easiest situation to land is "the road is closed, the route is fixed, the speed is low, and there is no one in the car", such as the mine car in the mining area and the unmanned logistics vehicle in the park; But the commercial value of this scene is limited, and its migration to other scenes is also limited.

AVP is in the middle-"the road is wide (but there are not many pedestrians in the parking lot, especially in the underground parking lot), the route is free, the speed is low, and there is no one in the car".

This is just a compromise, which is more malleable and has the following characteristics:

First, the legal obstacles are small-fast landing and large quantity.

Compared with the open and semi-open scenes in other cities, the legislative difficulty and ethical risk of low-speed unmanned AVP are obviously less.

There is no ethical problem of "protecting passengers or pedestrians" in this scene; Due to the low speed, there are fewer pedestrians entering and leaving the basement, and the risk of accidents is also small.

Tang Rui told Jiazi Guangnian that he believed that the fundamental reason why AVP could land faster was: "The laws and regulations are simpler, and the willingness of car manufacturers to order mass production will be stronger."

The direct commercial consequence of this is that AVP can land more quickly.

Second, it faces both VaaP and TaaS markets-it can make blood and has a large space.

In the VaaP market, Volkswagen, Daimler, BMW, Volvo and other companies have put the production models embedded in the AVP system on the agenda. The FAW AVP mass production order obtained this time also belongs to VaaP mode.

VaaP enables manufacturers of AVP to gain profits in a short time, and has certain self-hematopoietic ability.

At the same time, AVP has the potential to cut into the operational service, namely the TaaS market.

At present, the product form of AVP can directly cut into the car time-sharing operation service.

This mode is actually * * * enjoying the car version of the bicycle. However, compared with bicycles, taking and putting cars is more complicated. Time-sharing operators need to spend a lot of operating costs on dispatching vehicles, and the inconvenience of parking and picking up cars also affects the market acceptance of time-sharing leasing.

Therefore, time-sharing operators have a strong demand to adopt AVP system.

In the time-sharing mode, AVP vendors have the opportunity to participate in the operation and provide value-added services, such as providing data insight for the insurance industry, and adding precise marketing, destination discount information, advertising and other services to the passenger AVP control portal (such as mobile APP).

This is a business model with lighter function, greater marginal benefit and greater profit space than suppliers.

Third, AVP has the potential to cut into the "laurel stage" in the future.

As mentioned earlier, the investment in autonomous driving is vast, and the most exciting thing for everyone is the "laurel scene"-the high-speed autonomous driving scene of urban roads.

And AVP has the potential to cut into this scene.

This is because the technical requirements of AVP are very high.

This scene accords with the characteristics of open roads and free routes, which means that complete L4 autopilot technology is needed to ensure the large-scale commercial landing of AVP.

The main difference between AVP and autopilot laurel scene lies in:

The difference between the above two points is the bottleneck that restricts the landing of high-speed L4 automatic driving at present.

The former requires the use of laser radar with longer detection distance, but it is expensive and the cost remains high.

The latter needs a technological breakthrough. As John Crafts, CEO of Waymo, said, the popularization of autonomous driving will take decades, mainly because the technology has not broken through the highest level L5 that can drive in any weather and under any conditions. As long as this ability is not unlocked, the safety of autonomous driving cannot be really guaranteed.

In other words, there is no obvious gap between the companies that have the strength to formulate a reliable AVP scheme and those that aim at high-speed L4 autonomous driving.

In Tang Rui's words, anything is technically feasible. And those technologies that should be there are still waiting.

Secondly, compared with sub-tracks such as logistics vehicles and road trucks in the park, AVP is directly applied to passenger cars, which is more in line with the high-speed automatic driving scene of passenger cars in business and cooperation.

Becoming a low-speed L4 AVP supplier for the whole vehicle factory first, and then developing high-speed L4 automatic driving together may become a way to reach the laurel steadily.

"I think this is a strategic highland. If this is occupied, get on the expressway. This is very solid (solid foundation). " Tang Rui said, "In China, low-speed L4 autopilot may land first." .

The opportunity to catch AVP earlier before the end of 20 16 also benefited from the company's genes.

Throughout, it only represents a big faction of "car +AI" in the field of autonomous driving, and the other big faction is "AI+ car".

The so-called "car +AI" is to cut into automatic driving based on the automobile industry. The attempts of GM, Volkswagen and Toyota, and the attempts of Tier 1 (first-class suppliers) such as Bosch, ZF, Continental and Hyundai Mobius all belong to this school, which is characterized by product orientation.

Mu Zong was founded on 20 13, and it is also a "car school" autonomous driving company with a profound automobile background:

Tang Rui, the founder of Mu Zong, has 65,438+04 years of R&D management experience in the field of automotive semiconductors. The team also includes many world-class Tier 1 executives and technical experts, such as Chen, former chief engineer of Chery Automobile Engineering Research Institute. After the establishment of the 20 16 technical team, AI talents such as dorri, who was a senior manager of R&D in Zhuoran and CSR, were also introduced. Have a clear and sober understanding of the mass production process of the car factory and how to control the product cost.

"AI+ car" is based on AI technology, starting from the algorithm of autonomous driving, and gradually seeking commercial landing. Many companies with Internet technology background belong to the "AI+ car" school.

Its overseas representatives include companies such as Waymo, Cruise, Uber Autopilot, Auto X, and domestic representatives include Baidu Apollo and Pony.ai, Roadstar.ai, Wen Yuan Zhixing, etc. (the three companies mentioned later all have Baidu backgrounds).

But in the bleak environment of 20 18, the market is more concerned about who can live in the present.

In terms of "viability", the car +AI faction has certain advantages.

So is the development of holism.

20 13 to 20 17 is the first stage of the whole project, and the leap from start-up company to Tier 1 has been completed.

At this stage, the main products are L0 to L2 ADAS systems, which can realize the driver's auxiliary parking and driving functions for the last 3-5 meters.

By May of 20 17, we had established our own production lines in Xiamen and Shanghai, and obtained IATF 16949 certification (a production process quality system certification). From a second-tier supplier that mainly provides algorithms and solutions, it has been upgraded to Tier 1 that directly provides software and hardware products and services to car manufacturers. Cooperative car companies include Geely, Weimar and SAIC Chase.

At the same time, after the completion of the A round of financing, an autonomous driving team will be set up after the Spring Festival on 20 16. The biggest difference from the "AI+ car" school is that from the first day of autonomous driving, what I want is how to really mass-produce.

Tang Rui has three thoughts on how to realize the mass production of autonomous driving.

First of all, mass production must reach the "car regulation level" in quality.

Cars are travel tools, but from another perspective, they are also "killing machines". The reason why the automobile has become the industrial pearl is that any parts and systems used in the automobile have to be tested under various extreme conditions such as high and low temperature, falling, vibration and aging, so as to ensure safety and stability as much as possible.

This is a process that can't be accelerated. Gu Jianmin, CTO of Valeo China District, once said: "It usually takes five or seven years to develop a real car gauge from the scheme to the final mass production." As Tier 1, it has shipped 500,000 sets of car gauge products when it started to drive automatically, and has rich experience in R&D and mass production.

Second, while the quality is excellent, in order to increase the quantity, we must strictly control the cost.

This is also the reason why we focus on the low-speed L4 scene, because the low-speed scene does not need too long braking distance, does not need to use expensive components such as laser radar, and does not have too high requirements for the processing capacity of the computing platform.

Based on the visual perception scheme, with the low-cost millimeter-wave radar and the low-cost Qualcomm Xiaolong 820A platform, the cost of AVP is controlled very low. "We can manufacture the whole system, controller and all sensors within several thousand yuan," Vice President Chen told Jiazi Guangnian.

Third, mass production should also meet market demand.

Difficulties in parking, picking up cars and urban road congestion are special pain points in China's travel field. Therefore, Tang Rui judged that the high-speed scene in the United States may have certain advantages, but in China, the high probability is that the low-speed L4 autopilot will land first.

In this way, we started to develop AVP products in September of 20 16, and launched our own AVP 1.0 system at the end of 20 17 1.

Looking back on how to gain market opportunities, Tang Rui said: "This is a question of thinking mode. At that time, many people wanted to be a cool high-speed autonomous driving company and did not seriously think about what kind of autonomous driving is needed in the China market. "

When the commercial play of AVP is specifically started, the overall view also maintains the stability of "car +AI".

Stability is first manifested in capital lending. Because ADAS's product line is relatively mature, it is different from many large financing autonomous driving companies. Generally speaking, the hematopoietic capacity of ADAS is used to cover part of the cost of developing self-driving, so it is not so dependent on external capital.

In terms of expanding business model, we must first be a good supplier. From the beginning of 20 18, we started to push the AVP mass production order with the automobile factory, and completed the complicated process of technical exchange, SOR (requirement specification), bidding, winning the bid and factory audit, and finally got the letter of appointment from FAW at the end of the year.

In this process, the advantages of being a first-class supplier, which took more than three years to establish, began to appear, which were embodied in the good business relationship with the purchasing department of the car factory, the technical cooperation with the research institute of the car factory for many years and the tacit understanding with SQE (supplier quality engineer) in production and quality control.

At the same time, autonomous driving technology also gives Mu Zong the opportunity to jump out of the "first-class supplier" and upgrade its position in the industrial chain, that is, to cut into time-sharing operation services.

"I have a point that the L4 autopilot must be operable," Tang Rui said. "Time-sharing is a good opportunity we see. It can not only upgrade the traditional ADAS, but also bring new business models. "

At present, Mu Zong has made initial contact with the time-sharing platforms of several automobile manufacturers to explore the landing mode of AVP under the time-sharing mode.

After combing the five-year history, there is a clear upgrade path from assisted driving to automatic driving, which is not high-profile:

"From the perspective of the whole company, the strategy is very clear-how do we establish the leading position of autonomous driving from the incision of low speed and AVP? We have not changed this. " Tang Rui told Jiazi Lightyear.

On the eve of mass production of AVP, there are still some uncertainties in this market.

The biggest controversy is the connection direction of technical solutions:

At present, there is a spectrum distribution about how to do AVP. One end of the spectrum is "car-changing school" and the other end is "garage-changing school".

The overall goal is to change the automobile school and pursue "intelligent automobile", that is, to realize AVP without transforming the basement.

Tang Rui believes that the car reform plan has three major advantages:

First, it can better control the cost-the cost of rebuilding the basement is relatively high, and it limits the geographical scope of AVP use;

Second, Mu Zong itself has a strong cooperative relationship with the car factory, which can quickly mature the commercial characteristics of AVP through mass production and shipment. Rebuilding the garage involves cooperation with real estate developers and properties, and the market is fragmented and the regional attributes are too strong, which increases the difficulty of commercial landing;

Thirdly, the scheme of rebuilding the basement does not require high intelligence of the car, but the AVP realized by "rebuilding the car" is closer to the future high-speed L4 automatic driving scene, which is in line with the long-term development strategy.

"It would be better if the parking lot has communication and parking lot coordination, but if it is not modified, the car will have enough intelligence, just like people can find parking spaces in unfamiliar parking lots." Tang Rui said.

Horizon also belongs to the change school. Last year1October 28th, Horizon publicly tested its AVP project for the first time. However, Horizon's current scheme uses the WILDEN 16 line lidar (the official price is $3,999), and the cost needs to be further reduced.

Bosch, the world's largest first-class supplier, is the representative of "garage reformers". Through the transformation of the garage, different models can realize "automatic parking". The vehicle itself only needs basic functions such as electronic braking, automatic shifting, electronic power steering and remote interconnection, and does not need any environmental awareness function.

The advantage of this scheme is that the requirements for vehicles are low, but the current cost is high.

In the scheme jointly exhibited by Bosch and Daimler in China last year, they used an industrial single-line lidar mounted on a column. Based on the consideration of system redundancy, in practical application, it is possible to arrange 25 lidar for every 3 parking spaces on average, which is unrealistic in cost.

Holoparking launched by Heduo Technology at the end of last year is in the middle.

In the words of Ni Kai, the founder of Heduo Technology, it is a trinity scheme of "field end, high-precision map end and car end".

The difficulty of this scheme is the need for systematic research and development: "I need to do research and development on three terminals, and the face is relatively open, and I need to complete many technical modules at the same time." Ni Kai once said in an interview.

Whether it is "Car Reform School" or "Car Reform School", in the process of promoting the commercial landing of AVP, they all face the same problem, that is, they need to integrate resources from all parties and formulate a new way to distribute benefits in the industrial chain. This is because the application of AVP as L4 autopilot is divorced from the concept of single hardware module or software function. AVP technology providers are no longer simple suppliers of Tier 1, but more and more deeply involved in the whole life cycle of AVP "product+after-sales".

At present, there are six types of commercial roles involved in AVP scenes: AVP autopilot technology providers, vehicle manufacturers, time-sharing operators, high-precision map makers (the realization of autopilot technology requires a high-precision map of the basement), charging technology providers, and commercial real estate/property companies.

The problems that need to be clarified by all parties mainly lie in two aspects.

The first is the business model, which involves who will pay the bill in the end and how to distribute it upstream.

The payer is relatively clear-whether it is VaaP mode or TaaS mode, the final payer has a driver/passenger at the C end; Under the TaaS model, there can also be some "wool on pigs" income, such as marketing, advertising, insurance and other businesses combined with tourism services.

What all participants need to consider is how much the most terminal consumers are willing to spend on AVP system. Would you like to spend money on time or pay in one lump sum?

The upstream distribution model is even more uncertain. For a technology provider like Mu Zong, it may be necessary to stop the one-time charging mode of selling products and licenses in the past, and instead charge according to the number of calls, trying to cut into value-added services based on AVP; Graphic business and property also need to find a reasonable way of profit distribution in the new model.

Second, how to allocate power and responsibility: Who is responsible for the accident during autonomous parking? How should the insurance company pay?

"It must be the business model of service," Tang Rui firmly values the operation service combining AVP and time-sharing lease, but he also admits: "The specific distribution of benefits and responsibilities are still unclear."

But this "vague" state often means opportunities.

For the rising stars in this kind of automobile industry, they have the opportunity to gain the most favorable market position through a new round of reform.

"The whole market is fragmented, and no one has full voice. It is precisely because this matter is more complicated that a core technology supplier is needed to take the lead. " Tang Rui focuses on the current situation and possible positions.

At present, Mu Zong has discussed the new business rules with some automobile factories, graphic dealers, commercial real estate and other related actors.

Next, the long-term goal is to upgrade to high-speed L4 autopilot on the basis of AVP.

This step-by-step company has made a three-step plan.

The first step is to continue to work closely with the automobile factory to run AVP technology from 0 to 1.

In 20 18, Mu Zong obtained a large number of pre-research projects of domestic car companies. Mu Zong will announce more good news, including entering the supply chain of joint venture car companies, after the mass production of AVP orders for FAW Hongqi models.

Tang Rui's own judgment is that the key to running from 0 to 1 is that in 2020, when the number of vehicles carrying AVP reaches more than 654.38+million a year, AVP is considered to have passed the market test.

At the same time, Mu Zong began to lay out the second step, using AVP technology to improve the scheduling efficiency of time-sharing leasing service providers.

"I personally judge that time-sharing companies are more suitable to copy from 1 to N." Tang Rui told Jiazi Guangnian that such operators pay more attention to marketing and distribution, and it is difficult to meet their business development needs before their products are fully verified.

Therefore, the overall plan is to start talking with the time-sharing company about specific landing projects later-about the second half of 20 19.

Prior to this, Mu Zong will pass a large number of tests in a small area, visit all technical points of AVP and time-sharing lease, and then jointly operate with the big platform.

The third step is to truly enter the field of high-speed L4 autopilot in the future.

In this process, Mu Zong will still benefit from its Tier 1 status, and use ADAS and AVP product lines for self-hematopoiesis.

At the same time, when the technical conditions are more mature, the cost of sensors is reduced, and the legal and regulatory environment is more perfect, we will enter the "laurel scene" of high-speed L4 autonomous driving.

Tang Rui's judgment of "timing" is that the real competition of L4 products will take place in the next five to 10 years, which is a big enough opportunity and needs enough patience.

"The really exciting autopilot must be the service." Tang Rui said.

In the new era of the combination of autonomous driving and operation, we will not stay in the role of a traditional Tier 1, but consider more possibilities-for example, starting from the core components and deeply participating in a series of value-added services brought about by the change of travel mode.

"We hope to be the first company to prove that L4 can be truly commercialized in China." Tang Rui is a man who doesn't want to show his edge too much, but this time, he said he wanted to be the first.

First, who will spend it? Just like the development of autonomous driving, "haste makes waste", and you can't see the result in a hurry.

However, being the first to get the mass production order of FAW AVP has added confidence to Gaoda, and this company has won a favorable position. In the next two years, we will launch a sprint to the staged victory.