In recent years, the concept of “logistics automation” has frequently mentioned publically and has raise huge attention from the media, head capital and end customers. However, even in today, forklift trucks, tractors and other industrial vehicles, the most demanding material handling equipment in the market, still rely on massive manual operations in warehouses, factories, logistics parks, airports, wharfs and other logistics nodes. There are many “cold-down” thinking for the industry even under the hot trend of automation:is the final solution for logistics automation to “let vehicle conditioned to scenarios” or the vice versa? What is the future development direction and trend of driverless industrial vehicles?？
Recently,VisionNav Robotics won 36Kr 2020 China "Top 100 Venture Companies Concerned by Investors". 。About the future development and trend of logistics automation and driverless industrial vehicles, Dr. Li Luyang, CEO of VisionNav Robotics,accepted an exclusive interview with 36kr.
Vision is the inevitable choice for driverless industrial vehicles
Question 1: Why VisionNav Robotics choose vision as the core sensor for driverless industrial trucks? In the field of vision, what technical reserves and accumulations VisionNav Robotics have?
First of all, in terms of bionics, more than 80% of external information of humans and animals comes from their vision sensors, their eyes, and almost all the complicated operations of humans also rely on the feedback of the visual information. Secondly, the richness of information obtained by visual sensors is incomparable with any other sensors. Cameras (visual sensors) can collect color and texture information in the environment, and play a vital role in motion control and perception of autonomous vehicles and other types of mobile robots. Thirdly, the cost of industrial cameras is only 1/5 of industrial laser sensors with the same grades.
VisionNav Robotics is one of the incubating enterprises by The Chinese University of Hong Kong, with a deep accumulation in the technical field. The chairman of the company is Professor Liu Yunhui, an international expert in the field of robot and control, an IEEE Fellow, a pioneer in visual servo control, the Director of the CUHK T Stone Robotics Institute. The company’s core R&D team has been established for more than 9 years and has been always focusing on combining visual information with robot motion to achieve world-leading breakthroughs in two key technologies, namely visual servo control and visual environment perception of driverless industrial vehicles. There are 33 papers publishing in international top journals and conferences, and 45 approved invention patents/PCT .
"Let Vehicle Conditioned to Scenarios" Will Be the Eventual Solution for Logistics Automation in the Next Ten Years
Question 2: What are the application fields for driverless industrial vehicles? What are the critical performance indicators for customers?
Driverless industrial vehicle application fields are two major industries: factory logistics and warehousing logistics. Factory logistics includes 3C manufacturing, automobile & auto parts manufacturing, food & pharmaceuticals manufacturing, etc., while warehousing logistics includes e-commerce logistics, retail logistics, third-party logistics, etc.
There are mainly three performance indicators that customers pay attention to:
Firstly, safety and stability. Customer’s basic requirements are always the stable operation of the driverless industrial vehicles in 24/7 without any safety accident for 356 days of a year. Recently, driverless industrial vehicle of VisionNav Robotics has obtained CE certification in compliance with ISO 3691-4:2020 standard, which is also the first CE certificate awarded by TUV SÜD across the world that is compliance with the new standard. It shows that the safety performance of driverless industrial vehicles from VisionNav Robotics has reached the international level.
Secondly, efficiency, particularly the peak working efficiency of driverless industrial vehicles in complicated rigid demand scenarios. Logistics is determined by the business flow which is always changeable and unpredictable. And factory and warehousing logistics is not only affected by business flow, but also the unstable delivery time. Therefore, the requirement to driverless industrial vehicles for factory and warehousing logistics is very high as their peak work efficiency in complex environments
Thirdly, the change to the environments and process when using driverless industrial vehicles. Customers want as small change as possible, and there comes with three reasons:
（1）Cost and time for the change: many customers have low profit margins, or limit one-time investment and cannot afford any lockouts;
（2）Changing environmental reduces flexibility: when customer converting their business or moving their locations, one-time investment cost is substantial;
（3）Customer need to re-learn, adjust and adapt after the changeover of the original operation process, which brings barriers to replicate in large scale.
Question 3: In your opinion, how to classify existing logistics automation solutions? What will be the future development trend?
There are two categories for logistics automation solutions:
（1）Overthrow customer's original environments, processes, storage architectures, transportation modes and even management modes, what we call “let scenarios conditioned to vehicles”;
（2）Follow customer's original environments, processes and other inherent patterns, what we call “let vehicles conditioned to scenarios”.
VisionNav Robotics believes that the second solution will definitely stand out first according to the current situation of the logistics industry in China. Even after 10 years, it is also the eventual solution。
Reason is simple. In China, for both factory logistics and warehousing logistics, the entire logistics process does not produce high added value. In logistics nodes, both logistics and manufacturing companies need to achieve cost reduction and labor saving on the premise of high cost performance and low investment. Under this background, few customers will use the first solution, which means to invest heavy assets, unless for benchmarking projects but they only accounts for 1% of the entire scenarios (mostly are non-demanding scenarios). Therefore it is absolutely impossible to use the first solution and do large-scale replication in the remaining 99% of the demanding scenarios.
As for the second solution, under ideal conditions, customers don’t need to change or replace their original factory/ warehouse layout, passage width, racking, pallets, transportation processes, management systems, facilities including manual forklifts/tractors when automating. Only by adding a module, the manual forklift trucks, tractors, and other industrial vehicles can be retrofitted and transformed into autonomous vehicles, which can be put into use directly to replace labor. If they also can also operate in complicated demanding scenarios, with only 12-18 months return of invest, then it is very possible to scale (massively) and replicate (standardized)
As a result, VisionNav Robotics will firmly choose the second solution, to "let vehicles conditioned to scenarios". By continuously improving perception and motion control capabilities of driverless industrial vehicles in complicated and demanding logistics scenarios, let the performance of driverless industrial vehicles infinite close to manual operation without changing the customer’s environments, processes and modes, to better scale and replicate.
Factory logistics in manufacturing industry will be achieve logistics automation one step earlier, while combing vision technology and 5G technology is perfect
Question 4: In which industries driverless industrial vehicles will be the first to replace labor? In your opinion, how long does it take to replace labor?
I believe the use of driverless industrial vehicle will be more popular and faster in factory logistics. There are two important reasons:
（1）Processes, standardization and informatization in factory logistics are at a higher level than warehousing logistics. Although when comparing perception and motion control capabilities between driverless industrial vehicles and manual operation, there is still a large gap, the partial scale application can be applied in some controllable environments to form closed loop processes, thus the benefits of the customers are calculable;
（2）In China, manufacturing companies generally have higher profit margins and higher labor costs than warehousing logistics companies, therefore they can accept longer return of investment time for driverless industrial vehicles and other logistics robots, and with higher willingness to input.
What we cannot ignore for sure, is that the penetration rate of driverless industrial vehicles to the total industrial vehicle market is still less than 1% by the end of 2019. In other words, regardless of factory logistics or warehousing logistics, driverless industrial vehicles have not yet open up the rigid demanding market, or scale and replicate the rigid demanding scenarios “from 1 to N”. However, VisionNav Robotics believe that the day is about to come with the joint efforts from us and the industry colleague.
There are more than 20 doctors in VisionNav Robotics R&D team. We spend nine years to improve two things to promote the day to come in sooner time:
（1）Under complex and rigid demanding scenarios, to continuously shorten the gap, in particular the unit perception and motion control capabilities of driverless industrial vehicles and manual operation;
（2）To develop standard solutions and improve delivery capabilities for four types of complex rigid demanding scenarios, and make every endeavour to scale + replicate in specific scenarios within one year. Our goal is to achieve a 5% market penetration rate of driverless industrial vehicles to total industrial vehicles in China within three years.
Question 5: In the future, what changes will 5G and other new communication technologies bring to the entire industry? In the field of 5G technology, how VisionNav Robotics market itself?
The combination of vision and 5G technology in logistics automation is the general trend.
One of the main barriers to massively applying vision to automated guided vehicles is the lack of computing ability. There is no way to turn visual information into a vehicle’s motion control ability in a complicated environment at a low cost. 5G’s high transmission speed allows image information obtained from the vehicle end transmitted to the cloud or server in real-time for computing, and the calculation result back to vehicle end for execution.
In the near future, almost all driverless industrial vehicles including vision guided forklifts, tractors, and port cranes will not need to carry any computing systems, but only do three things: collecting sensor data, transmitting via 5G, and executing controller commands. The powerful cloud computing capabilities can make in-depth analysis with a large number of environmental image data containing rich color and texture information, convert them into perception content and motion control commands, and send back to the vehicle end for execution. It certainly will dramatically improve the motion capability of driverless industrial vehicles in complex environments, making it infinitely close to manual operation.
As an explorer of "Vision + 5G" in driverless industrial vehicle industry, VisionNav Robotics has cooperated with several industry leading customers for experimental projects. We look forward to VisionNav Robotics becoming one of the first enterprises in China to realize large-scale application and mass production of "Vision + 5G" driverless industrial vehicles.