What is flexible logistics? There are two things: fluctuant throughput self-adaption, scenarios self-adaption.
—— Andy (Luyang) Li
Chuangyebang 100 Future Business Summit and 2020 Annual Meeting was held during December 16-17, 2020. Guests discussed the new trend of entrepreneurship under the influence of Covid-19 during the year of 2020.
In the event, Jeffrey (Guangxi) Wang, Managing Director of Lenovo Capital had a conversation with Andy (Luyang) Li, Cofounder & CEO of VisionNav Robotics to discuss “innovation opportunities in vertical fields during the era of intelligent Internet”. Three main opinions are summarized as below:
1.The application of driverless technologies in logistics industry mainly rely on three core technologies:
the first thing is positioning, create an environmental map to let robots know its own position;
second is the perception, including the environment and the goods;
third is control, let the robots complete motions on the basis of positioning and perception.
2.Logistics is passive and decided by business flow. With the appearance of new business models,
business flows have changed its supply chain from "Push" to "Pull",
which has driven logistics to become more flexible.
3.What is flexible logistics? There are two things: fluctuant throughput self-adaption, scenarios self-adaption.
Topic of the Conversation: Innovation Opportunities in Vertical Fields during the Era of Intelligent Internet
Below transcript of the conversation is compiled by Chuangyebang:
Wang: I would like to thank Chuangyebang for giving us such opportunity for conversation. The year of 2020 is extraordinary. The epidemic has brought opportunities for many new industries. For example, the demand for unmanned industries has sprung up like mushrooms. First of all, let me briefly introduce Lenovo Capital. We focus on investing IT companies in early stage. Although we invested a wide range of companies, their industry applications is basically about “hard technology”, such as robots, AR, VR, Internet of Things, edge computing, cloud computing, big data, AI, etc. In the past few years, we have clearly felt a trend that, more and more technology startups are accelerating their application landing in the industry. Today we have invited Li Luyang, the co-founder and CEO of VisionNav Robotics, as one of the enterprise representatives to share his opinion.
Li: Thank you. I'm Li Luyang from VisionNav Robotics. Let me start with VisionNav’s vision: empower driverless industrial vehicles with vision technology, drive process of flexibility and automation in logistics nodes.
Let me explain. You can see VisionNav in two aspects: first, application of driverless technology in logistics segmentation; second, we are representative enterprise in flexible logistics automation. To make it simple, we are a logistics robot company. Our company is under 5 years old.
Wang: Many companies have been deploying industrial automation recently, including core-level component, robotic arms, control algorithms, and driverless technologies. In the past few years, the rise of AI vision and 5G gives us a lot of space to do things that were unimaginable, or that had no chance at all to be done at a suitable cost-performance. I would like to hear from you in detail, that how you to apply good technology and land those applications in industries, and the directions of your future development?
Li: Let me simply introduce the market first. Logistics is the third largest market in China, following real estate and automotive. So why logistics? I think the market of logistics market is super cool, here are two parts, "warehousing" and "distribution". Everyone is more familiar with distribution, for example, high-speed land transportation, last mile delivery, community new retail, etc., all belong to distribution. Whereas warehousing are the nodes in the logistics. Without the nodes, there is no way to distribute. The combination of warehousing and distribution will finally become a logistics network. For VisionNav, we pay more attention to warehousing rather than distribution.
What is the market share of node logistics? According to our preliminary statistics, the cost of internal logistics nodes is 1.9 trillion in China in 2019, including renting, material handling and other links.
What we do is to replace the manual operations with robots in these node logistics, such robots are also called driverless industrial vehicles. In node logistics like warehouses and factories, the main tool for material handling is forklift. People use forklifts to move and store pallets in rack shelves to complete the internal transportation inside the whole warehouse. And our job is to transform these industrial vehicles, mainly forklifts, into driverless vehicles to promote unmanned processes.
The application of driverless technologies in logistics industry mainly rely on three core technologies: the first thing is positioning, create an environmental map to let robots know its own position; second is the perception, including the environment and the goods; third is control, let the robots complete motions on the basis of positioning and perception.
How big is this market? The increment number for industrial vehicles that are produced in China and applied in logistics industry is 425,000 units in 2019, and the reserve number exceeds 3 million units, and the number of workers engaged in industrial vehicle driving is about 5 million. This is a relatively large market. However, the unmanned penetration rate is still relatively low. In the last year, the total increment number of driverless industrial vehicles nationwide was only 2,700 units. If it were not for pandemic, we would not receive many attention from the capital market. The pandemic has given this market a positive pushing.
Why is the penetration rate low? The gap between unmanned and manned is still very large, in terms of overall motion control, environmental perception, and positioning. Therefore, our company integrates visual AI and 5G technology when facing this problem, although it cannot immediately shorten the gap between robot operations and manual operations, but at least there is no difference now at the information acquisition side.
Jeffrey (Guangxi) Wang, Managing Director of Lenovo Capital
Wang: You have same opinion with Elon Musk.
Li: True, he said self-driving technology that not using vision is freaking stupid. In my opinion, there should be no difference on the information acquisition side, the larger gap is on the processing side.
When it comes to 5G, it is very valuable to combine it with vision technology. First of all, it is very challenging to process video stream on the vehicle side. Decoding the video stream, analyzing the information we want from it, and giving it semantics, is very stressful.
When increase the computing ability on the vehicle side, the costs also increase, then you will need to find another way to sell customers at low cost and in batches. However, with 5G, video streams can be collected in mass flows very quickly, edge computing can be done in the cloud, reducing the computing ability requirements of the vehicle side. In the future, 5G will be very empowering for vision, providing the possibility of greater computing power, making the efficiency of driverless industrial vehicles much closer to human.
Wang: from what you just mentioned, vision, AI, information acquisition, information processing and its timeliness, are closely related to end-side, edge-side and cloud-side in terms of low-latency applications. You also mentioned the low penetration rate, there may be many reasons, some of which are cognition differences, maybe also some special circumstances, thus not triggering the real demands. Can you give an example about how you do product design, at which cost-performance level, how to implement and how to gain benefit from scale economics, so that we can intuitively understand why now is a good timing for promoting driverless solutions integrate 5G, AI and vision?
Li: Let me explain with some figures: in the fields of driverless industrial vehicles (automated guided forklifts), the order we received this year accounted for about 10% of the total market in China. We may rank second or third. The leading companies in the industry, such as Sinopec, Nestle, Gree, Amway, are our direct end customers. There are about 20 top 500 companies like them are our customers. And we are also very happy to see that many of our customers are start duplicated from 1 to N. That means after trying from 0 to 1, they start to order from 1 to N. One of our customers, for example, place order for six time in a row, 4 of them are in this year, forming duplication from 1 to N.
What’s the reason behind? When we say cost-performance, performance comes first, it is the most important thing, that is, whether robot comparable to human? Is robot able to complete human’s work? For example, when automated guided forklifts to load and unload trucks, or to lift the goods to rack shelves, whether their efficiency comparable to human, such result in whether "robot replacing human" is feasible. If yes, next consideration comes to money (consider price), for example how long will it take for cost recovery. It seems during the epidemic, everyone keep silence for a quarter and did not open up much business, instead they stay in the company for R&D mainly. Maybe that’s why I can see in the second half of this year, the technology of this industry has been greatly improved, and the performance of machines has become closer to humans.
Andy (Luyang) Li, Cofounder & CEO of VisionNav Robotics
Wang: Can you share with us how you think of technology trends in the future. When we talked earlier you mentioned shallow-water and deep-water in the industry, and for deep-water applications, it always takes massive technical development to achieve something that other people cannot, which is often the keys to open up key customers. So how you see the trends of technology? How you define deep-water? And what kind of technological development can help you to achieve breakthroughs and keep the leading role?
Li: Logistics is passive and decided by business flow. With the appearance of new business models, business flows have changed its supply chain from "Push" to "Pull", which has driven logistics to become more flexible.
What is flexible logistics? There are two things:
First, fluctuant throughput self-adaption. Logistics is passive and unsteady. For example, if the truck is stuck in traffic jam for half an hour during trunk transportation, it will be half an hour late to arrive the warehouse, which will create an unexpected fluctuation, resulting in a sudden increase or decrease in the flow of people and logistics in the warehouse. Unmanned equipment should be able to adapt to such fluctuation, we call it fluctuant throughput self-adaption.
Second, scenarios self-adaption. There are many different types of goods with different sizes and in different batches in the warehouse or a storage area. If environment changes, it is very difficult for customers to re-purchase new equipment. Equipment should adapt to different changes. Even if the customers say that they are not using this warehouse anymore and moving to another place, equipment shall be able to used direct in the other place.
I think the most important thing for flexibility is the "performance" of the cost-performance, the flexibility equivalent to humans, while the most critical part is how to self-adapt throughputs while improving product performance.
Wang: In recent years, we have been continuously making strategic transformations, building large platforms to provide integrated solutions for invested companies. What do you see as the best opportunities for unmanned in some industries? Can you talk about your point of views on the industry?
Li: Thank you Director Wang. Since the investment half a year ago, VisionNav owes Lenovo Capital quite a lot advertising fees (with smiles), as we can often participate in various media activities to talk with many potential customers. Let me firstly talk about our relationship with Lenovo, and then talk about the industry.
First, there are a large number of associated companies with Lenovo. In the past six months, we have connected with Comac, Ansteel, Unilever, and Lenovo's own smart factories. There are a large number of related demands so that we can connect with many customers and resources.
Second, Lenovo invested a lot of excellent companies, many of which are related to logistics. A large number of companies invested by Lenovo are logistics solution providers and equipment suppliers.
Third, I found out that Lenovo has strong R&D capabilities and mature R&D background, which fits us very well. For example, Lenovo has made a lot of platform trials in 5G, cloud, cloud computing, and we can directly copy Lenovo’s solution and apply them to our customers.
When Lenovo Research working with Comac, cleansing robots and deicing robots have been developed for aircraft. These all well fit with VisionNav’s technology.
Another example is that in terms of core sensors, we surprisingly find out that Lenovo has mad very mature camera which we can use. So I think that Lenovo can use its own platform and technical capabilities to do some integration, gathering companies like us to serve related companies with Lenovo's integration capabilities.
Wang: Now there are a lot of scattered demands. Large and medium-sized companies hope to have integrated solutions. With this opportunity, we hope VisionNav Robotics, our invested companies and all the other companies present here, can make contribution to the industry together. Welcome everyone to further cooperate with us.
About VisionNav Robotics
Established in 2016, VisionNav Robotics (Shenzhen) Co., Ltd. is a global leading company in driverless industrial vehicles. Focuses on combining vision technology with mobile robots, VisionNav provides driverless industrial vehicles and driverless logistics solutions for manufacturing logistics and warehousing logistics. At present, it has developed and produced a variety of highly flexible products such as driverless industrial vehicle modules, indoor automated forklifts, outdoor automated tractors, etc., breaking through rigid-demand scenarios including 9.4 meters high storage, 2.0 meters narrow aisle storage, automated truck loading and unloading, multi-layer mobile racks stacking, which are widely used in e-commerce logistics, third-party logistics, 3C, pharmaceutical, FMCG, food, automobile manufacturing and other fields. VisionNav Robotics are committed to promoting large-scale replication of driverless industrial vehicles in complex rigid demand scenarios, and realizing technological innovation and industrial application in China’s driverless logistics industry.