Ali has many animals that can make blood by themselves, but Dharma Academy is the biggest exception. Three years ago, Ma Yun said that “will live two years longer than the group” of the science club. After the early publicity stage of overexposure, it is destined to enter a long and silent sweeping monk mode.
After all, as long as “basic research” is involved, the time dimension for measurable results is calculated in terms of “years” and “ten years”. Even Ali internally joked that the work of the Dharma Academy is divided into “primary hardship”, “intermediate hardship” and “long-term hardship”.
For example, if AI technology-related laboratories are considered “junior difficult”, then those whose difficulty has reached the Boss level are, of course, quantum computing laboratories that “may not survive in 10 years.”
So in the past three years, every time we have the opportunity to chat with the scientists of the Dharma Academy, we will definitely make a joke:
“How did Ali set KPI for you?”
In fact, Ali does not care how to get the code on the hands of these scientists to the front line of the huge business. On the contrary, no company is willing to have the same attributes as a university campus, allowing scientists to live in their own ivory tower.
From the generally acclaimed “Polatao” on Taobao on mobile phones to the emergence of “Plattou” after the acquisition of Zhongtian Micro Semiconductors in 2018, Dharma Academy has gradually changed the “and current” set when it was first established. There is a seclusion mode where the R&D system remains independent and focuses on medium and long-term technology R&D.
The current sweeping monks, in addition to doing research, also shoulder the responsibility of helping Ali raise, domesticate and hatch animals.
In this Yunqi Conference, after Brother Pingtou, the Bodhidharma Academy “born” the second animal that desperately needs to earn money to return blood-a donkey.
At the Yunqi Conference, Zhang Jianfeng, Dean of Alibaba Dharma Academy, released this machine
The rookie brother is a donkey
Little donkey. Except that the name is very lack of technological temperament, the application scenarios it represents are ready to be seen from the meaning of the word-since it is to help humans carry goods, it must be related to logistics.
In fact, this product, which is similar to the “autonomous driving logistics car”, is not a new thing. Because three years ago, the rookie ET laboratory also introduced a similar species. So you should be able to guess that behind the new product is a new round of structural adjustments experienced by Ali Dharma Academy and Cainiao——
The autonomous driving laboratory of Dharma Academy was integrated with Cainiao’s previous unmanned logistics vehicle research and development department. The latter was completely merged from the Cainiao system after being separated from the Cainiao system, forming a new team of about 300 people, and the name is Donkey”.
On the one hand, the goal of the Autonomous Driving Laboratory established three years ago is to explore the commercialization direction of related technologies such as autonomous driving and robotics. The Cainiao ET Laboratory is specifically responsible for the development of autonomous driving logistics vehicles. There are certain directions between the two. Coincidence.
According to a source to Huxiong, the group has made comprehensive considerations from avoiding duplication of vehicle manufacturing and streamlining business directions, and finally decided to complete the merger of the two departments within a few months. In May of this year, the self-driving logistics trolley running in the Jiang’an Community of Sichuan University actually had the words “Cainiao ET Laboratory” attached to it.
Dr. Gang Wang, a scientist at the Autopilot Laboratory, has become the new leader of this team.In an exclusive interview with Tiger Sniff, he explained that the brand “Xiaomanlu” will become the second technology company incubated from Dharma Academy after Pingtou. Cainiao has become one of the biggest customers of Xiaomanlu Robot Company.
Indeed, from the perspective of commercial landing, given that the promotion model of Cainiao Station is relatively mature in many areas, it is more difficult for the self-driving logistics trolley to be promoted to various closed parks as the “standard” of the Cainiao Station than it is to find a better one. Father Fang had much less difficulty in negotiating cooperation.
This is a typical business strategy of Ali, which is to fight a cooperative battle through the end retail ecology established by itself.
In a retail ecosystem filled with Taobao, Tmall, Ele.me, Cainiao, AutoNavi Maps and other related companies, Xiaomanlu will become a small part of the hardware elements of this ecosystem-building machines suitable for various scenarios , And then connect these products to each specific business unit. In this way, there is also the possibility of realizing mass production and returning to the cost.
“It is normal for Dharma Academy to burn money, but once you have a corporate identity, you have to make your own blood. The relatively easy way to break through is to use existing businesses to do supporting and business extensions.”
Wang Gang told Hu Xiu that many universities, including Shanghai Jiaotong University, Sichuan University, and Zhejiang University, have set up Cainiao Stations, and self-driving delivery cars have also taken this “opportunity” to be included in Cainiao’s service design process. It is the same cabinet system and Alipay payment system.
“Part of the cost of the daily maintenance of the machine can be saved, because the staff at the rookie station can share part of the daily cleaning tasks.”
In addition, just as express delivery is paid per order, logistics trolleys have also set up the same payment model. Students from Jiang’an Community of Sichuan University told Huxi that the use of the logistics car is free only for the first time, and that you will have to scan the code to pay for it later.
From a technical point of view, compared to the unmanned vehicle scene on urban roads, in theory, the safety and speed requirements of autonomous driving in a closed park are greatly reduced, which increases the feasibility of commercialization.Of course, the complexity of the scene is different, and the difficulty of handling the problems of passenger cars, trucks and unmanned delivery vehicles cannot be compared.
“There is no lane line in this scene, and there are no clear restrictions on traffic participants. Bicycles, skateboards and motorcycles are all passing through the school. So the environment we face is in a dynamic change.”
Chen Junbo, technical director of Xiaomanlu, pointed out in an interview with Tiger Sniff that since it is to deliver express delivery, the importance of “delivery efficiency” is actually higher than that of purely demanding cargo safety, which increases the requirements for the response speed of logistics vehicles. .
When they started the test, they found that the robot would often be “frozen” because it felt that the environment was too complex and any action might bring safety risks, so it stayed in place.
“In the university’s first season, all kinds of luggage and parcels will be placed on the ground at will, and there will be construction scenes. A pit may be dug at will without obvious safety signs. There will be some difficulties in recognition accuracy. So from the school gate When driving to a dormitory area a few kilometers away, we must develop a comprehensive logistics vehicle that can overcome diverse obstacles.”
In the beginning of the school season, such stalls abound
According to the feedback Huxian collected from the students of Sichuan University, many people have indeed seen the blue self-driving logistics trolley parked at the door of their dormitory in recent days, and some people have indeed taken out the package from it.
A boy who has used a logistics car told us that because he had been using the rookie station before, he tried an unmanned delivery vehicle. “I was playing games in the morning and didn’t want to go to the door to pick up the logistics vehicle. It was indeed delivered downstairs within the prescribed time.”
From a cost point of view, according to the information obtained by Tiger Sniff, the cost of the unmanned logistics car launched by the Cainiao ET Laboratory two years ago is about 200,000. Although Ali did not disclose the cost of the latest machine, Wang Gang said that there will be substantial reduce.
What’s interesting is that not only the ecology of the rookie station provides more sales channels for Xiaomanlu, but in the past few days in Hangzhou, I found that many mid-to-high-end residential buildings have also displayed “Damo House” and “black”. The concept of science and technology, and the unmanned delivery vehicles deployed in the community are one of the “strong proofs” given by real estate developers.
A person familiar with real estate in Hangzhou told Huxi that, just as the surroundings of Marvel movies usually sell well, the contribution of Dharma Academy and scientists has exceeded their own business:
“Dharma Academy is actually not just a pure basic research institution. Its brand and influence are actually gradually becoming a soft culture and penetrating many other industries, becoming a powerful marketing tool.”
Is it a gimmick or a real need?
The track of “end logistics distribution” has relied on the unmanned delivery vehicles of Ali Cainiao, Meituan, and JD.com three years ago on the wave of information offensives, and it has become one of the fastest scenes of autonomous driving in the capital. One”.
But if you zoom in from farther, you will find that there are many unsolved problems.
First of all, success is speed, and failure is speed.
The speed of this kind of trolley is usually 10~20km/h. Although the speed reduction is “a huge benefit” to ensure the safety of pedestrians and the real-time performance of the system, when the speed is reduced to a bottom line, the delivery efficiency will be compromised. .
A student who has used unmanned delivery vehicles at Sichuan University told Huxi that the space in the trolley is quite large and can hold many large packages, but if it is not particularly difficult to get packages, she would not choose unmanned vehicles. Because “the speed is too slow, you can catch up with this car by walking faster.”
Secondly, for the unmanned delivery of the last mile, the payment model requires further planning by the enterprise. The second time I have to pay more than 1 yuan, many student parties feel that there is not much need to use unmanned vehicles to pick up packages. Especially for girls who like to go shopping with e-commerce and have a lot of packages and scattered, the cost is actually not small.
Some people even joked: “Cyclists are reluctant to buy a monthly pass, let alone pay for express delivery for only a few kilometers each time.”
Therefore, for the above two reasons, there are not many students who really frequently use delivery vehicles.
In response to student feedback obtained by Tiger Sniff, the Xiaomanlu team explained that “Sichuan University is currently a trial of the logistics vehicle distribution payment model, and will cancel the payment in the near future and implement a completely free system.”
“Except for Sichuan University, all of our trial operations in other universities and communities are free. In fact, Xiaomanlu has not determined the final business model, and all things to make money are still being explored.”
Third, the spaciousness and flatness of the ground of university campuses and high-end communities are the basic prerequisites for the smooth operation of unmanned vehicles. For many old communities with high and low thresholds, uneven roads, and narrow roads with little space, there is limited copy space for unmanned delivery scenarios.
Fourth, take the Jiang’an Community of Sichuan University as an example. There are more than 20,000 students living in Jiang’an. If conditions such as cost and delivery efficiency are all possible, a small car that can carry 18 packages is definitely far from enough, but if the number exceeds a certain number, new requirements for campus traffic and community capacity may be put forward.
How to cooperate between multiple vehicles to minimize the time cost, while maximizing the load of each vehicle as much as possible (more than 18 pieces will make the package easier to be confused?) is the future need of the Xiaomanlu team Ask the Ele.me scheduling system team for advice.
During the Double Eleven period, the express delivery ceremony at Sichuan University Jiang’an Campus
The action route of the self-driving car can be seen clearly from the two places with red circles in the lower left corner of the figure. The car needs to plan 16 dormitory buildings.
Plan of Sichuan University Jiang’an Community
Fifth, Xiaomanlu once assured Tiger Sniff that “there will be no problems with mass production capacity”, but if the model is replicated in more communities and campuses in the future, or even out of a rookie to provide the same products to more logistics companies, Will put forward higher requirements for process design, after-sales technical support team and sales team.
“Although Cainiao is our customer, it does not mean that our stuff is self-digested internally. We still have to open up more customers in other industries, including how to promote our robot platform. We have to sell hardware and also To export technology.” Wang Gang’s idea is somewhat similar to Alibaba Cloud’s technology mid-station model.
However, in the terminal logistics and distribution market, the output effect of autonomous driving technology capabilities is never determined by technology.
Although the “contradiction between the aging population and the scale of one billion parcels” has always been raised by technical experts on various occasions, at this stage, there are actually few large companies that seem to have strong demand for unmanned delivery. .On the one hand, it is still unclear how much cost can be saved, on the other hand, after all, stuffing parcels into the car still needs people.
in fact,What logistics companies and takeaway platforms need more is the “human-machine cooperation” mode of courier brothers and autonomous driving, And this is another comprehensive problem that is as difficult as climbing.
Xiaomanju is indeed more commercially viable than self-driving cars on urban roads, but whether it can achieve a large enough sales scale and obtain considerable returns is a new development after a group of scientists and engineers left the Dharma Institute. Journey.
Alibaba Cloud New User Welfare Session