What is the principle of zero grounding integration?

When will robots replace high-altitude Spider-Man?

Source

What is the principle of zero grounding integration?

When will robots replace high-altitude Spider-Man?

Source | Founder Park

High-altitude robots are gradually replacing the "Spider-Man" working at high altitude.

Different from the well-known service robots serving human beings and industrial robots used in production environments, special robots were born to liberate human beings from various high-risk environments. Common special robots include underwater robots, ground mobile robots and high-altitude robots. Replacing humans with high-altitude robots is also an effective way to reduce the risk of high-altitude work and improve efficiency.

At present, high-altitude robots have been used in many fields such as industry, agriculture and forestry. Shihe Robot, a representative enterprise in China, specializes in ship derusting, chemical storage tank anticorrosion, thermal power boiler inspection and building external wall cleaning, and provides professional and complete solutions, which has occupied the first place in the domestic ship derusting market.

Dr. Xu Huaying is a guest in the geek park "Founder Park" "Reconstruction" column | Source: live screenshot

What is the difference between the core technology of high-altitude robot and industrial robot? How to standardize the solution in the scene landing? And, what are the technical problems faced by high-altitude robots?

On May 3 1 day, the founder of Shihe Robot &; Dr. Xu Huayang, CEO, visited the geek park to "rebuild" the live broadcast room and talked about the technical status and future market prospects of high-altitude robots.

0 1

Progress in material technology

Let the high-altitude robot

realize

Founder Pu: What problems does the facade robot mainly solve?

Xu Huaying: The facade robot, as its name implies, is a robot that walks on the facade, which is also commonly known as a high-altitude robot. The application scenarios are very extensive. What is common is the external walls of buildings in cities, which can replace the "Spider-Man" that we often see. In addition to this civil scene, there are also rich aerial work scenes in industry: ship rust removal, storage tank corrosion prevention, boiler inspection and bridge inspection.

We divide the function of the facade robot into three links: construction, inspection and repair. Creation is the application of welding and polishing like shipyard and tank factory; Inspection is the application of inspection and flaw detection like boilers and bridges; Repair is the application of rust removal, painting and external wall cleaning. In short, the facade robot can replace the traditional manual work in the three links of construction, inspection and maintenance in the high-altitude facade scene.

Founder Park: The common UAV uses multi-rotors to fight gravity. How do high-altitude robots solve this problem?

Xu Huaxuan: There are still some differences between high-altitude robots and drones working in facade scenes. UAV is mainly equipped with detection equipment for macro detection, such as aerial photography or mapping. Of course, some non-contact operations can also be carried out, such as spraying pesticides. Drone. High-altitude robots have stronger load capacity and can be used for close observation. It can be equipped with EMAT electromagnetic ultrasonic thickness gauge and even eddy current thickness gauge for very accurate detection. Derusting requires a very high bearing capacity, not only to bear ultra-high pressure water pipes and cables, but also to bear the recoil force of ultra-high pressure water jet, about tens to hundreds of kilograms. Stronger load capacity, higher positioning accuracy and longer endurance are the main differences between high-altitude robots and drones.

How to fight gravity? The principle is different in different scenes. The mainstream is magnetic adsorption in industrial scenes and negative pressure adsorption on non-steel surfaces.

The commonly used material in industry is steel, and magnetic attraction can be used for steel, so that adsorption can be realized on the surface of steel. But the difficulty lies in the good coordination of suction, gravity and driving force. If the suction is too large, it can't be used flexibly, and if the suction is too small, it is easy to slip and crash.

If the surface is non-steel, for example, the external wall of a building is a glass curtain wall, it can be solved by negative pressure adsorption. For tubular or columnar objects, friction can be generated by friction, hug or support, so that they can crawl directly along the pipeline.

Fang Zhengpu: It sounds like it should be able to resist gravity, effectively carry loads and move flexibly. What kind of technological progress makes this easier to achieve?

Xu Huayang: In the final analysis, it is a problem of statics and dynamics. If the robot is on a flat steel plate and bears a fixed load, it is a statics problem. Such a wall-climbing robot existed 20 years ago, and the mobile chassis can walk on a relatively flat steel plate with magnets.

However, such an ideal environment does not exist in both industrial and civil scenes. The real challenge is to adapt to the curved surface in order to crawl and cross the weld in various positions of the ship. Especially for thermal power boilers, the water wall structure is complex and uneven, so it must have the ability of real-time magnetic field perception and adaptation. This was not available to wall-climbing robots many years ago.

The core technology is mainly two points: the progress of NdFeB materials, especially the improvement of magnetic energy density. The higher the magnetic energy density, the stronger the magnetic attraction of magnets with the same weight, which can bring greater safety margin. If the safety margin is large, its bearing capacity can be guaranteed. With the progress of rare earth materials industry in recent years, the magnetic energy density of NdFeB materials has been improved. NdFeB materials belong to rare earth materials, and China is an exporter of 90% rare earth materials in the world. We are the world leader in permanent magnet material technology.

The second is the increase of the power density of the driving motor. The greater the power density, the greater the driving force under the condition of relatively compact volume or weight, which also means better safety margin.

These two are the core innovations of key materials and key components. Combining the breakthrough and innovation of the two, the wall-climbing robot talents can fly over the eaves and walk over the wall, which really has better adaptability in complex industrial application scenarios.

02

The goal of starting a business is to become

"DJI" of high-altitude robot orbit

Founder Pu: What prompted you to have the idea of starting a business?

Xu Huayang: I graduated from School of Mechanical Engineering, Central South University. I entered department of mechanical engineering, Tsinghua University on 20 1 1 to study for master's degree, doctor's degree and doctor's degree.

I have the idea of starting a business because I saw the biography of Jobs on 20 14. Jobs went from a poor young man to the CEO of the most valuable company in the world. A person has realized this value through his own efforts, which has brought great changes and influences to the PC, smart phone, animation and other industries in the whole world. Jobs' experience inspired my passion and motivation to start a business and gave birth to the idea of starting a business.

20 15 During my PhD, I tried to start an intelligent hardware business with my classmates. After the project started for more than a month, it obtained the enlightenment investment. We quickly moved from demonstration to engineering prototype, and the project was acquired a year later. 20 17 I started my business full-time after my Ph.D., and my orientation at that time was high-altitude robots.

Founder Pu: Why did you choose the field of high-altitude robots at that time?

Xu Huaxuan: For professional reasons, I have been paying attention to robots. I also like reading science fiction, especially Asimov's robot series, which has a great influence on me. What is the vision of the robot described by Asimov? There are many robots in every family. They help to serve tea and water, do laundry and cook, act as personal assistants at work, and also act as business assistants when traveling. It is the state in which robots and humans were born.

Singularity theory holds that with the development of information technology in human society, the performance of artificial intelligence will develop exponentially, and many viewpoints believe that strong artificial intelligence will appear in 2050. Looking forward to the arrival of the era of strong artificial intelligence, robots can enter thousands of households and empower thousands of industries. This industry may be bigger than the automobile and mobile phone industries.

I am very firm in the general direction of robots. However, it is necessary to find a pragmatic entry point and gradually accumulate and precipitate brands, funds and talents in the industry.

I am personally interested in mobile robots. Moving is equivalent to having legs, adding some executive devices is equivalent to having hands, adding sensory systems to have eyes, ears, nose and tongue, and then having brains and cerebellums to have human embryonic forms. So movement is a very important feature, and we wanted to find the application scene of mobile robot at that time.

Emotionally, when we are looking for business opportunities, it is May 1 day to 1 1 month, which happens to be the peak period of external wall cleaning. We can see Spider-Man working on the wall every day. At that time, we felt that such traditional and inefficient jobs should no longer exist. We also talked with the bosses and workers of more than a dozen Spider-Man engineering teams in Beijing, and found that everyone is also looking forward to robot products, because it is increasingly difficult to recruit workers in the industry and the cost is getting higher and higher.

"Spider-Man" for cleaning exterior walls | Source: Unsplash

Intellectually, we can't just do this scene. In the past, mobile robots did three scenes: sea, land and air. We feel that the scene of exterior wall cleaning belongs to a very independent and brand-new opportunity. At that time, it was positioned as an elevation scene, and the whole category was called a high-altitude robot. We thought there should be many other application opportunities here, so we went to the industrial industry to understand the situation. Soon, we found that the wind power scene was very suitable, so we tried it for the first time, but it didn't work. However, through such an attempt, it is found that there is still demand for robots in other scenes such as ships, chemicals, energy and bridges. With pain points and just need.

Combined with such perceptual knowledge and rational analysis, we chose the high-altitude robot as the starting point for starting a business. We hope to stay on this track for 8 to 10 years and become the "DJI" of the world's high-altitude robot track.

Founder Pu: Regarding the facade robot, have you done any market research or analyzed the current situation of foreign technology industry?

Xu Huaying: The easiest way is to analyze competing products at home and abroad and study their technical routes, as well as the industries and customers they are targeting. The first company analyzed is called ICM, an international climbing machine, and their first scene is wind power. We also imitated the direction of the wind. In addition, we also studied the thermal power detection robot of Gecko Robotics, the chemical tank detection and flaw detection of inspection robot company of GE, and the ship rust removal robot of VertiDrive company of the Netherlands. We classified these according to industries and functions, and found out the direction that is easier to land in China.

At the earliest, the wind power tower was cleaned, then the second project was the pipeline climbing robot, the third was the high-speed railway station column site, and the fourth began to try to remove rust from ships. In addition to our own active exploration, after years of precipitation, some projects have actively found us, because we also received customized orders for survival in the early days, such as Dongfang Thermal Power in Shijiazhuang, including Zhejiang Special Inspection Institute, which is also our customer.

In this way, through our own judgment and passive order taking, the basic demand pool at the top of the funnel is formed. Then in the process of doing it, judge its technical feasibility and commercial value, then further productize it, and then continue to polish and iterate from generation to generation.

Founder Pu: How did you put some early single projects into production?

Xu Huaying: The first product we landed was ship derusting, which took almost two and a half years to realize.

I first heard this demand from a veteran in the robot industry. He met such a customer's demand and found it technically difficult, so he suggested that we do it. There are hundreds of thousands of ships in the world and there is a great demand for derusting. However, the technical difficulty is also great, the hull is smooth and equipped with ultra-high pressure water pipes, so the robot is easy to fall down.

First of all, we made a survey of competing products and found that there are robots in this direction abroad. We think this scheme is basically feasible. There are also universities in China that have done research before and have some technical data. On the basis of their technical route, we further improve the magnetic attraction density of magnets and the adaptability of curved surfaces. Because we have done wind power projects before, we use surface adaptation technology here. Ultra-high pressure derusting has been found by some friends, and the key supporting facilities in the upper reaches have been gradually built.

I quickly made a prototype and found it feasible after testing with iron plate, so I began to look for early angel customers. In this process, we are mainly looking for innovative customers who are willing to adopt early adopters. Because the effect of prototype is often not very good. Through the recommendation of acquaintances, I found two customers, in Jiangyin and Zhoushan. Constantly testing in these two sites, at least there are actual scenarios, and the entire R&D chain is closed and can be iterated.

The first small batch will be started at the end of 20 19. After the application went online, the product problems were more fully exposed, and a small version iteration was made in time. V20 version has been carried out in small batches, and its consistency and stability are qualified, so it can be commercialized. We choose the way of service operation to let manufacturers make full use of it. It took about half a year, and our ship repairing capacity exceeded 100. It broke through 100 working objects, which basically proved the feasibility of business and technology.

Ship Derusting Operation | Source: Shihe Robot

After the release of the V30 version, the stability is better, the usability is greatly improved, and the auxiliary driving functions such as automatic cruise, one-button lane change and automatic obstacle avoidance are added, so that novices can get started quickly and have the potential for large-scale promotion. Subsequently, we started to open sales outlets in Qingdao, Ningbo, Guangzhou and Shenzhen to promote our products on a large scale.

Product and marketing are a cyclical process. The better the product, the wider the application and more and more customers. Because more and more customers use the products and get more and more feedback, the products will get better and better. These two iterations have a certain Matthew effect, because the scene is very barrier.

Founder Pu: What is the problem solved by automatic navigation of ship robots and what is the core principle?

Xu Huaying: Robot operation can be divided into two modes: manual remote control and autonomous operation. Autonomous operation, path planning and automatic navigation can be realized on the surface of rules. There will be some complicated areas at the stern and bow, so it is necessary to switch to manual operation mode and do some assistance through sensor cooperation.

We call the navigation technology of ground mobile robot an inversion scheme. From the inside out, it is equivalent to people wearing glasses and identifying their own positioning by combining the surrounding point clouds or image data. For appearance, we call this technology an outside-in scheme. The robot walks on the wall, and a pan-tilt camera is set in the third view. Because our positioning and navigation don't need absolute position, only relative position. For a ship, if it is converted into a rectangular frame in the image, the robot only needs to traverse from the upper left corner to the lower right corner of the rectangular frame, and the rust removal is completed without telling the absolute position. Based on the image from the third perspective and the data from the robot's own multi-sensors, 3-5 cm precision positioning can be achieved to meet the needs of the scene.

Fangzheng Pu: 100 What is the workload of ship cleaning?

Xu Huaying: A normal cargo ship of 6,543,800 tons, about 300 meters long. In the past, it took 30 or 40 units to work at the same time, about 50-60 people, and the same number was needed for night shift. If robots cooperate with humans, 20 robots, 20 people during the day and 20 people at night, will be cleaned in about three days and three nights. It can greatly save costs and significantly improve efficiency.

Founder Park: What is your current revenue scale?

Xu Huaying: The specific value is not convenient to disclose, so you can simply say it.

The annual output value of China ship repairing industry is about 30 billion. The rust removal of the outer plate is 2-3 billion/year, and the amount of rust removal in the inner cabin is similar. There are nearly 30 head shipyards in China, accounting for about 80% of the domestic ship repair industry. More than 20 of them were directly or indirectly served by us.

So far, our robots have accumulated more than 500 ships. In this sub-industry, it should be the highest number of ship repairs in the world. The derusting area of ships is estimated to be over 600,000 square meters, and the market share is definitely the first.

Founder Pu: How big is the market for ship cleaning?

Xu Huayang: More than 70% of the global logistics is by sea. At present, there should be 6.5438+0.2 million cargo ships. During the epidemic, shipping is also more popular. We believe that in the future 100-200 years, shipping will remain the most cost-effective logistics mode.

Ocean voyage needs to solve the problem of ship plate corrosion. At present, the shipping industry needs to stop for maintenance twice every five years. As China is the largest importer and exporter in the world, ship repair is generally carried out near the import and export area. Half of the ships in the world are repaired every year, accounting for about 20%-30% of the global output value.

We chose the boat scene for a reason. First of all, the market space is large, which is convenient for early attempts. Second, the high concentration of customers is conducive to promotion. Third, the scene is highly standardized. Although there are different ship types, they are basically similar scenes relative to the size of the robot, and the technology is very versatile.

03

Shihe robot technology

Advantages and business model

Founder Park: Apart from the early entry time, what are the core technical advantages of Shihe Robot?

Xu Huaxuan: Shihe has been developing a mobile platform for high-altitude robots from both industrial and civil scenarios. The core technology has four directions:

Firstly, the magnetic adsorption mobile platform, the matching of adsorption force and driving force, from the design analysis, optimization and control process are very particular. In the design process, it is necessary to consider the adsorption force and driving force under extreme conditions, and there must be a safety margin to meet such a scenario. Secondly, what you see is what you get, and the designed magnetic attraction module is equivalent to a real experiment. Then in the process of motion control, it is best to sense the change of magnetic attraction in real time and adjust it in time through the change of magnetic field, which has better adaptability. Finally, the load-bearing ratio is large, the surface adaptability is strong and the obstacle-crossing performance is good.

The second is a non-sustainable facade robot driven by rope and negative pressure. Compared with magnetic attraction, negative pressure has the problem that it can't cross obstacles, so combining negative pressure with ropes to adapt to glass curtain walls and building external walls also involves the positioning of building external walls and crossing obstacles.

The third core technology is the positioning and navigation of the facade robot, that is, the outside-in multi-sensor fusion facade positioning technology just mentioned, which is unique in the whole industry. Our friends, including foreign manufacturers, also adopt the mode of pure remote control operation. Next, we also plan to have more than one person, because one person can operate multiple machines with independent operation functions, further improving the return on investment for customers.

The fourth is the process package of the industry, which is an integrated system of software and hardware. The load is its hardware. There are also some process parameters that are part of the software. For example, when ultra-high pressure water jet is used for derusting, the parameters of water jet will affect the efficiency of derusting. These parameters are related to the pressure and flow rate of water machine, the arrangement, number and target distance of nozzles. All these need to be accumulated many times to find out the rules, in order to give full play to the highest efficiency of robots and water machines in various scenarios. Parameters and laws may be more important than hardware in essence, because hardware is easy to design and imitate, but parameters can only be summarized and precipitated by themselves.

In other words, the chassis is serialized and standardized, so it can be used in various scenarios. As for the assembly on the chassis, we also have two ideas. The first is to give priority to cooperation and development with industry partners. In the past, various industries had traditional handheld technologies and corresponding tools, giving priority to cooperation and integration. If we can't adapt to the robot well, we will consider turning to self-research.

Founder Pu: Technically, what practical problems have Shihe robots mainly solved?

Xu Huaying: Take the ship as an example. The linearity of the ship is very complicated. For example, at the ball nose, the radius of curvature is about 1 m. When the robot is working, it is easy to crash if it can't perceive the changes of magnetic field and curved surface in time.

Detection of Thermal Power Boiler | Source: Shihe Robot

Complex scenes include thermal power boilers. The burner is several hundred meters high, and the inside is a side-by-side tubular structure, which needs ash removal, glue removal and flaw detection. There is also a scene that crosses the folding angle, and the magnetic field will suddenly change. Only three companies in the world can realize the corner crossing function in the scene of water wall of thermal power boiler. Gecko Robot and Inspection Robotics, a wholly-owned subsidiary of GE, are in the United States, while Shihe Robot has only one in China.

In addition, the elevation positioning navigation and process package mentioned above are also brand-new technical advantages.

Fangzheng Garden: Building exterior wall cleaning is also the scene you have to solve. What are the key technologies?

Xu Huaying: Exterior wall cleaning should be the biggest subdivision scene in the trajectory of aerial work. The outer wall area of buildings in first-and second-tier cities alone is about 654.38+0 billion square meters, and the annual cleaning fee is about 20-30 billion. Considering that the frequency of cleaning up first-tier cities is higher, the annual cost is more than 30 billion.

The difficulty of this scene is that making a single explosive is too complicated and nonstandard. Finally, we must combine the actual situation of different scenes to make a series of products. There are series of product components, the main modules are common, and there is potential for scale.

However, there are also ways to divide scenes, using STP methodology. We divide the whole building exterior wall into six categories. The first category is the simplest pure glass curtain wall. The second is the curtain wall with only mullions, because there are only mullions, one column at a time, and there is no need to cross obstacles. The third is a curtain wall with only horizontal frames, which can walk sideways without crossing obstacles. The division of the first three categories is mainly to try not to cross obstacles, because it is too difficult and inefficient to cross obstacles. The fourth category is defined as having both vertical frames and horizontal frames, which need to cross obstacles. The fifth category is that the overall shape is relatively regular, but there will be some complex structures on the surface. The sixth category is alien, such as the bird's nest or "big underpants".

At present, we mainly focus on the first four categories, and we will not do the fifth and sixth categories for the time being. In the future, we may cut into the fifth and sixth categories by making customized special planes for big B customers. In the short term, it is still aimed at the front comparison head, and the stock is relatively large.

Founder Pu: It sounds like the high-altitude robot product is a very complicated system. Is the upstream and downstream technology and equipment purchased and integrated, or do you do the whole chain yourself?

Xu Huaying: Many parts of the whole industrial chain are common, such as motors, reducers and drivers. Some special magnets and ultra-high pressure water pipes are special suppliers. We have an assembly plant in Fangshan District, Beijing, which mainly does testing, assembly and inspection, and the supply chain is still done by upstream and downstream partners.

The industrial chain is also relatively mature. Motors and reducers are made in China, and many chips are general-purpose chips, which are not affected by trade wars. Magnets are China's export materials and are not affected. Therefore, the whole industrial chain is relatively complete.

Founder: How do you plan your future business model?

Xu Huaying: We call it the sales +RaaS (robot as a service) model, and RaaS charges according to the service. Going to sea is very important, and the concentration of domestic customers is high, so the direct sales model is adopted, and the channel sales model is more used abroad.

The overseas strategy we understand may be globalization+localization. Globalization must be done, but the people who do it must be local. In the short term, it is still necessary to focus on channels. Because overseas resources are not enough, overseas friends are also selling or renting. Leasing equipment is still a mature business model overseas.

Founder Pu: If you have enough capital support, what kind of technical route will you take?

Xu Huaying: The first is the richness of product lines, including vertically integrated product lines in various industries. For example, we have derusted the outer plate of the ship and then derusted the inner cabin. The next step is spraying. After repairing the ship, go to build a ship and infiltrate it, and then go to sea.

The energy industry, such as testing the productization of exterior wall series, has been set as a benchmark by gecko robots abroad, with great commercial certainty. Domestic thermal power boilers have also entered the later operation and maintenance market, and the demand is increasing. Finally, there are six scenes of civil exterior wall cleaning, and the first four scenes will be commercialized as soon as possible.

The second is the global layout. We are now a mature product, a product that goes to sea. Ship derusting robots and chemical anti-corrosion robots have been polished in China for about three years and are mature, which can accelerate the layout and expansion overseas.

The third is the domestic business of RaaS, or it is hoped that it can quickly integrate with large funds, seize relatively high-quality users, establish cooperative relations with engineering service companies in advance, and form a * * * ecological environment with shipyards or chemical plants for a long time.

These can be done for at least 6-8 years. It is necessary to make a series of product lines, the landing and globalization of RaaS business, and lead the whole trajectory of high-altitude robots. After 6-8 years, Shihe hopes to grow into a super unicorn enterprise on this track and become a global multi-billion dollar company.

* Head shot Source: Shihe Robot