Running discussion on SP, page-20345

Logistics market starts to use Driveless electric Truck?!


Driverless electric truck starts deliveries on Swedish public road
Esha Vaish

JONKOPING, Sweden (Reuters) - Resembling the helmet of a Star Wars stormtrooper, a driverless electric truck began daily freight deliveries on a public road in Sweden on Wednesday, in what developer Einride and logistics customer DB Schenker described as a world first.

Robert Falck, the CEO of Swedish start-up Einride, said the company was in partnership talks with major suppliers to help scale production and deliver orders, and the firm did not rule out future tie-ups with large truckmakers.

"This public road permit is a major milestone ... and it is a step to commercializing autonomous technology on roads," the former Volvo executive told Reuters.



The logistics auto-driving war begins! 5 major forces 30+ players robbed 9 trillion market
2019.02.02 09:30:29 Creations

Text/car stuff

Source: Car stuff (ID: chedongxi)

In 2018, Google Waymo took the lead in wielding the banner of commercialization of manned autopilots, but after a while, the crowd found that it was still "trial operation."

At the same time, the automatic driving of the logistics without the human light carrier is in the dark. In China, facing the 9 trillion logistics market, the parties tried to re-segment this high-priced cake with autonomous driving technology.

Logistics auto-driving startups, truck OEMs, suppliers, Internet giants, and logistics operators have all entered the market. More than 30 companies in the industrial chain are attacking the city.



Five major forces to transform the 9 trillion market

In fact, the logistics industry's demand for autonomous driving is much stronger than the demand for automatic driving by passenger cars or large passenger cars.

In China, logistics costs account for more than 10% of GDP. Although this figure has been decreasing year after year, it still reached 14.5% in 2017.

In 2017, China's logistics costs were 12 trillion. Among the 12 trillion yuan, road transportation accounts for 76%, and simple conversion, domestic logistics on the ground, has formed a market of 9 trillion.

The 9 trillion core driving force is 14 million cargo trucks that are mainly active in medium and long-distance transportation and intercity transportation, and nearly 30 million vans, tricycles and two-wheeled vehicles that are mainly active in urban transportation, express delivery and delivery. Electric cars, motorcycles.

Behind the 44 million vehicles, of course, the “drivers” are bigger than the 44 million figures (the logistics industry usually stops, especially medium and large trucks with more than one driver).

Even with the most conservative number of employees (44 million) plus the salary standard two years ago - the average annual salary of the logistics industry in 2017 is less than 80,000 yuan, the domestic logistics industry has only paid 3.5 trillion in the human salary of logistics. yuan.

Specific to a single company, Xia Huaxia, general manager of the US Mission Unmanned Business Unit, revealed to the car that the annual labor cost of a full-time takeaway rider of the US Mission was about 100,000 yuan.

Ma Yiren, CEO of the trunk logistics startup Qichen Technology, also said that the average cost of a trunk logistics truck driver reached 180,000 a year.


Drivers and takeaways, couriers always have to rest, while autonomous vehicles can be around the clock, and the potential for theoretical efficiency is much greater than humans.

The high total labor cost (because of the lack of insurance and benefits, the above-mentioned 3.5 trillion is only part of it) provides the most basic driving force for the introduction of autonomous driving in the logistics industry.

On the other hand, the serious security problems facing the logistics industry have also made the value of autonomous driving increasingly prominent.

Give two typical data. According to the data of the Traffic Management Bureau of the Ministry of Public Security, in 2016, China’s freight vehicles accounted for only 12% of the national motor vehicles, but they produced 48% of accident deaths.

In 2017, the Shanghai express delivery and take-away industry involved 117 traffic accidents, 9 people were killed and 134 people were injured.

Under the dual pressure of cost and safety, the domestic logistics industry is increasingly calling for autonomous driving. All players gather to respond and form the five major forces of automatic logistics (in terms of computing units and sensors, logistics vehicles and passengers). The car does not have specificity and is not listed here):

More than ten start-ups such as Zhijia, Tucson Future, Mainline Technology, AutoX, Feibu Technology, Xijing Technology, CIDI, etc., took advantage of the artificial intelligence dividend and cut into the traditional field of logistics, trying to be in the automation innovation of the logistics industry. Become a new generation of key technology suppliers.

ICT giants, whether it is Baidu, which is one step ahead in the development of autonomous driving, or Ali, Jingdong, Meituan, which has its own logistics and distribution business, and even Huawei, which starts from the communication point of view, will drive automatically as its future business. one. When these giants master the auto-driving technology for the logistics industry, they can export solutions to the industry and open up new sources of income; retreat to optimize their own logistics and distribution business, and reduce costs.

And logistics companies that rely on logistics as their main business, as well as truck fleet operators - whether it is SF, which has a market value of more than 300 billion yuan, or the Man Group, which has a valuation of 10 billion US dollars, are eager to reshape the industry. Put a flag on the technical high ground and firmly hold the territory that was originally your own.

The commercial vehicle mainframe factory that produces trucks has the same anxiety as the passenger car giants – the logistics industry is the largest market for their products, and the implementation of autonomous driving in the logistics industry will change the competitive landscape of products and the survival they rely on. business model. The main engine manufacturers of commercial vehicles are also afraid to become foundries.


▲ Volvo abroad opened the development of self-driving trucks in 2012
For parts suppliers of commercial vehicles, especially trucks, they see new opportunities brought about by product upgrades in the industry, such as various types of autopilot sensors, as well as commercial vehicle line-controlled chassis that have been difficult to promote on trucks. With the demand for automatic driving of logistics, a larger market has been acquired.

The above five categories of forces have maintained a competitive and cooperative relationship with each other because of different motives, forming a primary logistics autopilot ecosystem.

2

Snatch three business scenarios

In 2018, after years of development, autonomous driving gradually diverged from general technical development and discussion, and became a field scene card position war. This is no exception in the field of automated logistics.

It has been observed that the auto-driving related enterprises are currently competing for the three core scenarios of “one high (speed) two low (speed)”.

“One high” refers to the mainline logistics scene of high-speed sports; “two lows” is the last one-mile delivery and closed scene of low-speed driving.

1. Mainline logistics automatic driving: the most players, the most competitive

First analyze the logistics scene from the main line.

In the three major scenarios, the main line logistics has the largest overall capacity and the highest output value, and it is considered to be relatively easy to implement technically because the vehicle is mainly operated in a closed scene.

Therefore, this scene has attracted the largest number of players to "clam". Among them, start-ups are especially numerous, and every start-up that attempts to make gold on this track has its own technical skills.

In 2016, Tucson, who was transferred to the automatic driving development of trucks, has a luxurious return-to-automatic driving research and development team. CTO Hou Xiaodi, COO Hao Jianan, and chief scientist Wang Naiyan have all been the academic stars of Silicon Valley.

Since the mainline logistics is basically in a high-speed scene and the vehicle runs at a high speed, the environment sensing range of the automatic driving system is very high. Tucson has developed a sensing technology that utilizes a telephoto lens to achieve a maximum detection distance of 1000 meters. Tucson, who has added this technology, opened a commercial trial operation in Arizona, USA, and each self-driving truck earned $3,300 a week.


▲ Tucson unmanned truck in the United States
Similar to Tucson's future, the founder of Plus.ai is also a returnee background. Its CEO Liu Wanqian and CTO Zheng Zheng are still Stanford's classmates. Zhijia's unique technique in technology is to select some of the key pixels for fusion of the "medium fusion" sensing technology, to balance the automatic driving perception and component cost performance requirements.


▲ Zhijia Technology unmanned truck tested in North America
In the past year, Zhijia and FAW, Manbang set up a self-driving truck "software technology - vehicle platform - operation management" commercialized iron triangle.

In the case of Feibu Technology, which was born later than Tucson and Zhijia, there was a former chief scientist He Xiaofei sitting in the town. After less than one year after its establishment, Feibu Technology allowed the self-driving truck to run on the road.

Unlike other startups, FEP is choosing one of the core hardware of self-driving autopilots, the computing chip. Its first generation of autopilot dedicated computing chip (ASIC) Phoenix 100 has been streamed last year.


▲Flight driving technology and auto-driving trucks built by Debon Logistics
In the logistics trunk scene, it was late, and the company was established in early 2018. In order to deal with the post-production, it adopted the game of alliance. Last year, with the help of Shanghai Auto City Innovation Port, it established a similar Apollo's self-driving trucks for data sharing, technical exchanges, and regulatory pushes.

It is also worth mentioning that the G7, which is backed by the company, is one of the largest commercial vehicle management platforms in China. The number of commercial vehicles connected to the platform has exceeded 800,000 in 2018. This means that Autotech's autonomous driving technology has a huge data source and potential application targets.

In addition, startups directly received blessings from Internet giants. For example, the Changsha Intelligent Research Institute (CIDI), which was established in Hunan, was an autopilot project hatched by Dazhan, chairman of the Board of Directors of Hong Kong University of Science and Technology, and was awarded the investment of Baidu Venture Capital in February 2018. At the technology conference of Baidu Apollo, CIDI created a high-speed auto-driving heavy-duty truck, which was also featured as a typical example of Baidu's Apollo ecosystem.

Although the logistics trunk auto-driving scene is concentrated, the technology reuse is high, and the output value is large, it is an ideal business scenario, but the startup company does not have a good technology or a backing, it is difficult to emerge in it - because ICT (information communication Technology) The giants naturally cannot miss this good business.

Baidu, Ali, Jingdong and even Huawei have laid down their influence among them.

Baidu has cut into this scenario by investing in CIDI, and will not repeat them.

Ali has a strong vertical integration motivation for automatic logistics driving due to the huge logistics business derived from home appliances. Alibaba's autonomous driving program is led by Wang Gang, chief scientist of the Artificial Intelligence Laboratory, directly targeting the L4 level. Although Ali's self-driving trucks have never been officially unveiled, their recruitment information has been clearly understood - "to build a smart logistics vehicle."

Contrary to Ali’s unmanned team, Jingdong, who has its own logistics business, showed off its own L4 unmanned truck at the 618 conference last year and said that its unmanned heavy truck has been completed in the US for thousands of hours. Test.


▲Jingdong L4 unmanned heavy truck
Huawei, which does not do logistics, has developed a technology for automatic driving of formations from the perspective of communication technology enabling automatic driving. A number of self-driving trucks form a tightly-formed fleet, and exchange the data status of each vehicle in real time under the low-latency communication technology provided by Huawei to achieve the unification of multi-vehicle control strategies. This technology is used even if no one The driving skills are immature, and it can also rely on the human driving of the head car, and the rear drones follow the form to reduce manual participation.


Of course, professional logistics companies will naturally not relax their attention to new technologies.

With a total investment of 10 billion US dollars, it has formed a strategic alliance with FAW Jiefang and Zhijia, and will let unmanned logistics trucks on the road in 2021;

Tencent’s G7 has personally hatched the start-up company of the company to promote the development of automatic logistics technology for trunk logistics;

Compared with traditional Suning, its logistics business also cooperates with Zhijia to launch L4-level unmanned heavy trucks;

Shiqiao Group, which has been deeply involved in the logistics field, has introduced Baidu in its latest financing of 1 billion yuan, and plans to apply Baidu's automatic driving technology to the trunk logistics scene.

Even SF, which is known for its excellent confidentiality work, has also demonstrated its ambition to build autonomous vehicles through its “Automated Driving Simulation R&D Engineer Recruitment” in its recruitment information.

The automatic trunking of logistics trunks is also the main battlefield of commercial vehicle OEMs. Their products are headed for the logistics market. However, since the basic ADAS is limited in the promotion of trucks, the automatic driving accumulation of the main engine manufacturing plant is generally not deep, and usually takes the form of cooperation as soon as it is on the field.

In the early days, there was a situation in which a mainframe factory was bound to a startup company, such as Tucson-Shaanqi and SAIC-Xijing. However, since 2018, the OEMs have generally become more open-minded and have begun to adopt a more open cooperation model. FAW Liberation announced this year that it will build a "Columbus" open system and build intelligent logistics with a number of startups and operators.

2, the last one kilometer distribution: the battlefield of the Internet giants

When the autopilot scene switched to the last mile of distribution, the advantages of the Internet giants were highlighted – the two major businesses that make up the scene: takeaways and express delivery, all of which are dominated by Internet companies. The main companies active in this scene include the US Mission, Ali Cabbage, Jingdong and so on.

The last mile of delivery and delivery is highly dependent on labor, but there are typical structural manpower problems – especially during late nights, meals, off-hours, and holidays or special weather. Whether it is to meet the peak demand of sporadic or to cut labor costs, Meituan, Ali, and Jingdong have real demand for unmanned vehicles.

However, because the vehicles that are delivered at the end are often light and compact, and the shape is more like a wheeled robot, the traditional truck manufacturers and suppliers basically do not have this business. The lack of supply-side professional companies makes it possible to grasp the demand side. Internet giants have directly become the "talkers" of this scene.

In 2018, the Internet giant's unmanned distribution cars are on the road.

In April, the ET Logistics Laboratory of the Alibaba Logistics Company launched the “Kipras” with a load of 100 kg and a speed of 15 km per hour.

In June, during the Jingdong 618 event, Jingdong’s third-generation unmanned logistics vehicle was delivered on the road. Coincidentally, its running speed was also 15 kilometers per hour.


In July, the US group's take-out unmanned delivery car was officially released, and the unmanned distribution open platform was also launched, attracting Segway, Neolithic, Zhixing and AutoX, Roadstar.ai and other autopilot startups.


▲ US Mission unmanned delivery car
Internet companies with talents and user advantages have actually mastered the technology of the last mile of distribution autopilot and even the standard discourse right, and are gradually defining this subdivision of logistics autopilot.

Although the logistics giant SF has a huge demand for end distribution, its automation is currently mainly in the UAV, so there is less sound.

Under the background that the Internet giant controls users and directly builds up their own technology, the pattern of the last mile of distribution autopilot “giant with startups” has become more obvious. This scene is the main business startup and eventually acquired. More likely.

3, closed scene: business scattered flowers

The closed scene of automatic driving of logistics is not a unified scene, but a general term - as long as it is isolated from the outside traffic, it can be called a closed scene, and such a scene is diverse.

From the current situation of the automatic driving of logistics in the closed scene, the business of subdivisions such as ports, airports, industrial parks and mines has been excavated.

The closed scenes of these subdivisions are technically broken, and the output value of a single scene is low. Large companies are often reluctant to do such dirty work, but instead provide some startups with inspection, upgrade technology and even revenue to survive. Soil.

Founded in 2015, Xijing Technology developed early brain chips, but in 2017, after receiving feedback from customers, Xijing Technology began to develop the research and development of port autopilot technology, and with the support of SAIC Hongyan and other enterprises, it was successful. In 2018, the port autopilot demonstration of container trucks was completed in Zhuhai. Now port autopilot has become one of the main businesses of this once brain-like chip company.


▲Xijing Technology Port Area Unmanned Container Truck
The main line technology founded by Zhang Tianlei, who was born in the background of the Baidu unmanned vehicle project, also drove the auto-driving container truck into Tianjin Port early.

The Tucson future mentioned above also entered the port area for automatic driving in April 2018.

In addition to its passenger car autopilot business, Bianshi Technology has also opened a technical team for unmanned special vehicles for transporting goods at the airport. The unmanned vehicles built have been put into operation at an international airport.

Huituo Intelligent, which was invented by Qingdao Intelligent Industry Technology Research Institute, reached an in-depth cooperation with Linggong Heavy Machinery in September 2018 to provide automatic driving technical support for the latter's unmanned mining machine operating in the mine.

3

Automatic landing before the passenger car? Solve these two problems first

Due to the strong demand for automatic driving of logistics, and the running scenes and ethical issues involved “look” more simply than manned automatic driving, the industry has always said that logistics automatic driving is expected to land earlier than manned automatic driving.

This view is theoretically correct. However, in the context of China's autonomous driving industry, the landing of automated logistics needs to solve two crucial problems.

1. The overall cost is too high to recover technical investment

Compared with passenger cars, the automatic driving of unmanned vehicles is often more costly, especially in the two areas of trunk logistics and closed scenes. Because the car body is usually large and requires more sensors, it is a large sensor. Cost, logistics, and autonomous vehicles will increase dramatically.

An executive of a self-driving truck startup told the car that it only operated in a closed scene, and its truck sensor cost reached 200,000.

If you want to run in an open scenario, the sensors of the unmanned trucks will have to be upgraded and the cost will double.

One thing to mention is that in China, the cost of self-driving trucks is even more unacceptable – in the US, the average salary of truck drivers is four times that of China, even if the technical cost of self-driving trucks is slightly higher, it is easier to run. Win high labor costs.

In China, the average monthly salary of truck drivers is only 6,000 yuan. Once the cost of auto-driving is slightly higher, truck drivers can still be in a superior position in terms of cost.

Therefore, if the automatic driving of the logistics is more popular than the automatic driving of the passenger car, it is necessary to take a greater cost reduction.

2. The problem of popularization of line control is to be solved

Similarly, in the field of commercial vehicles (trucks), the necessary remote control system for automatic driving is also very poor. Even some technical solutions have to use the motor or the robot arm to directly hit the steering wheel or step on the throttle instead of passing the electric signal. Direct control of steering and acceleration, braking.

In China, there has not been a large-scale supplier of truck chassis control, and the control of logistics auto-driving vehicles has been controlled.

Domestic commercial vehicle OEMs are currently providing some of the open control protocols, which are carried out by downstream companies. There is no unified line-controlled chassis supply, which also raises the realization cost and difficulty of logistics automatic driving to a certain extent.

At present, the chassis control of commercial vehicles mainly relies on the foreign Tier-1, especially the industry leader WABCO, and Germany's Knorr.

However, domestic startups have begun to appear in this area, such as Zhongyunzhi, which will build a line-controlled chassis for commercial vehicles as a special entrepreneurial project.

Conclusion: Logistics Autopilot expects more cooperation

From the summary of the car things, logistics auto-driving is indeed as the practitioners say, there is a stronger demand than the passenger car field and more clear goals and business scenarios.

Relatively independent of passenger car automatic driving, automatic logistics has gradually become a self-contained, forming an industrial microclimate in the country.

In the face of 9 trillion industrial opportunities, whether it is a startup, a host factory, or a technology giant, they have high expectations and tangible investment in growing into a mature industry.

However, the realization of automated logistics also faces greater cost pressures and greater infrastructure difficulties. Rather than saying that this is a competition, it is better to look at it from the perspective of cooperation among all parties in the industry. “No company can make autonomous driving alone”, and this sentence is also true in the field of logistics.