After five years of building a business in the robot controller field, Zhang Xiaolong still firmly believes that the robot industry will usher in a glorious era for domestic brands, just as the smartphone industry did. And his goal is clear: to lead iNexBot to become the "Android of the robot industry".
Over the past year, against the backdrop of the Sino-US trade war, demand in the automotive and 3C sectors declined, China's manufacturing industry faced enormous difficulties, and the machine tool and robot markets were deeply depressed. Domestic robots are going through accelerated consolidation; the market has shifted from growth to replacement, and competition is unprecedentedly fierce.
However, the changing external environment doesn't seem to have affected iNexBot — the past year was actually the fastest-growing period in the company's history. According to Zhang Xiaolong, in 2019 iNexBot received orders for more than 3,000 robots, accumulated over 200 customers, and its products covered welding, cutting, spray painting, polishing, stamping, injection molding, loading/unloading and other applications — becoming the sole supplier for many robot body manufacturers.
What boosted Zhang Xiaolong's confidence even more: that year, iNexBot beat many industry veterans and listed competitors in the market. "In the next two years, we hope to fully surpass international brands such as KEBA," Zhang Xiaolong said.
Robots are saying goodbye to wild growth and entering an era of intensive cultivation
"Robots are the hope of manufacturing development worldwide — and we're just getting started."
Over the past 10 years, China, as the world's largest robot market, has grown nearly tenfold, with total capacity currently approaching 200,000 units and expected to grow to the million-unit scale in the coming years. On the other hand, China has roughly 90 million workers; even replacing just 1% of the workforce per year would require nearly a million new robots, and on a global scale the ceiling is several times higher.
Starting in 2012, China's robot industry entered a period of rapid growth that lasted until 2017 — the era of "wild growth", when policy and capital together fueled the industry's boom.
Now, with subsidies reduced and capital returning to rationality, the robot industry has ended its pampered dividend era. The competitive bar has risen quickly; storytelling and pie-in-the-sky promises no longer work. Only those with the ability to survive in the market can meet the challenges, and any pseudo-technology product that cannot be quickly deployed to solve real production and service applications will be rapidly eliminated by the market.
After 2018, with the gradual maturation of the domestic supply chain combined with broader social changes, China's robot industry entered a new phase: saying goodbye to wild growth and beginning an era of intensive cultivation.
Domestic robot manufacturers face two mountains: automotive and parts manufacturing, and 3C electronics. Together these two fields consume 70% of industrial robots, yet domestic robots have an extremely low market share there and can only find applications in the long-tail markets of small and medium-sized enterprises.
In Zhang Xiaolong's view, robot localization means that robot-intensive industries such as automotive and 3C will inevitably become the decisive battleground between domestic and imported robots — because these are the fields that truly determine the future of the domestic robot industry, and domestic manufacturers must prepare well for the decisive battle.
Therefore, domestic robots must both accumulate experience in long-tail markets and focus on developing new robot application areas, while actively pursuing import substitution in major application fields such as automotive and 3C. Today, domestic robots have outstanding technical service and price advantages over imported robots in developing new applications across many long-tail markets — this is the current foundation of domestic robot manufacturers.
In the long run, Chinese manufacturing can capture more than 70% market share in any competitive field. Today, domestic robots — especially vertical articulated robots — hold less than 20% market share. That is enormous market space and opportunity.
"2020 has the potential to be the first year of the domestic robot rise." In a robot market full of sixpences, Zhang Xiaolong is a rare soul who sees the moon.
One reality that must be faced: today's weak domestic robot body manufacturers still rely heavily on imported components — control systems, servo motors and reducers are still predominantly imported. "Assembling from imported parts, domestic robot body manufacturers can never defeat international giants like ABB," Zhang Xiaolong said.
Looking at the controller market, KEBA has undoubtedly become the market leader, with industry-recognized stability and control precision. Its conveyor tracking and DELTA processes are extremely mature, but welding, spray painting and other applications are largely missing. In China, companies that purely make controllers are rare — currently only a handful, such as Advantech Baoyuan, Googol, Huacheng Controls and iNexBot.
Meanwhile, some robot body manufacturers with a certain technical base hope to follow the Apple route — developing their own control systems and vertically integrating the industry chain. But Zhang Xiaolong believes they won't skip a single pitfall.
"From an integrator's perspective, a large number of incompatible self-developed systems ('little Apples') create serious fragmentation of control systems, further eroding the competitiveness of fragile domestic robot products. Every new control system pushed to market brings serious promotion, training and learning costs. Moreover, once a body manufacturer runs into operational problems, it also creates long-term maintenance issues for control." Zhang Xiaolong said.
In the era of intensive cultivation, channels and processes become the competitiveness and barriers of robot body applications, while real-time operating systems, bus technology, robot algorithms and the like are all common industry infrastructure technologies that cannot reflect the core competitiveness of a robot application enterprise.
Zhang Xiaolong believes that developing robots doesn't require starting from operating systems and algorithms. It's like opening a restaurant — you don't need to grow your own vegetables; developing an Android app doesn't require starting with CPUs and 5G chips. Phone manufacturers like Xiaomi and Huawei also rely heavily on ODM suppliers for software and hardware solutions.
In fact, many domestic companies originally started development on secondary platforms like CoDeSys, because there was no suitable motion control development platform in China. Now iNexBot's NexDroid — the only domestically developed robot control secondary development platform in China — is gradually taking root in the market. Globally, apart from CoDeSys and KW-Software, which implement robot control secondary development platforms based on soft PLC architectures, it is the only robot control secondary development platform provided in the form of C++ APIs.
Zhang Xiaolong said: "The NexDroidOS platform is designed to go head-to-head with CoDeSys, so that future robot companies don't need to write robot control algorithms to develop robots — they get all the software and algorithm fundamentals, making robot development simpler."
"I want the market 3–5 years from now"
Four years after moving from the smartphone industry into the industrial robot industry, Zhang Xiaolong clearly feels the difference between the B2B and B2C models: B2B is much slower. In his own words: in 4 years, a B2C company could have died two or three times over, but in B2B, it's just getting started.
For iNexBot, which is just getting started, talking about platforms now may seem premature — after all, most domestic controller companies are still struggling at the edge of survival.
A couple of years ago, another company raised this idea, but it gradually faded away — presumably because they saw how hard this path is. Even Canonop, which once held the largest market share in the domestic robot controller field, has turned to making robot bodies. So why does iNexBot, a rising star, persist in this?
Zhang Xiaolong said: "Building a platform serves two purposes: first, to reassure major customers that we have no intention of making robot bodies; second, to make integrators understand that they should build processes with us."
In the past two years, some body manufacturers have extended into core components, and quite a few core component companies have moved into bodies. One negative effect of the complex coopetition in the robot industry: everyone has become paranoid, because you never know when your partner will become your competitor.
"iNexBot is positioned as a motion control technology provider. We will firmly play the role of a core component supplier for robots and will never enter robot body manufacturing." Facing today's competitive market, Zhang Xiaolong is very clear about iNexBot's positioning.
What iNexBot wants to do is leverage the support and user base of body customers, adopt an open and interconnected attitude, provide highly intelligent solutions, hide the complex algorithms and process logic, offer users the simplest, easiest-to-use operation solutions, and lower the barrier to robot applications.
"Robot applications will be everywhere in the future, which requires a large number of companies to participate — but the barrier to robot development is too high," Zhang Xiaolong said.
Through sustained effort, iNexBot has gradually broken into some mid-sized robot body companies. Zhang Xiaolong found that what he needs to overcome is not just technical problems, but more importantly how to dispel customers' concerns.
"They don't worry about our product capability — they believe that given time and customer applications, we'll definitely succeed. What they worry about most is whether I'll die, or whether I'll be 'kept' by a body manufacturer." That's exactly why Zhang Xiaolong keeps emphasizing that iNexBot is a neutral player.
"We want to be the Android of the robot industry, solving the common foundational technologies of the industry so that the vast majority of domestic robot bodies run on our system." Positioned as a platform product, Zhang Xiaolong's ultimate goal is to realize the vision of "making robots simpler and robots everywhere".
In Zhang Xiaolong's view, the robot control system is the carrier of processes and the tool of human-machine interaction — essentially operating-system-level software, naturally monopolistic, and the system with the largest user base will ultimately win.
"I want the market 3–5 years from now," Zhang Xiaolong said.
Core Foundation
In "The Soul of a New Machine", Tracy Kidder vividly wrote: "Technology is the common soul of the people who develop it. Technology is fascinating, but what is even more fascinating is the developer's obsession with technology."
Zhang Xiaolong deeply relates to this, so when asked what iNexBot's core foundation is, he answered without hesitation: iNexBot's people.
It is said that iNexBot's technical staff are all experts from the internet, software and robotics industries — a group of idealists who pursue ultimate technology. "We firmly believe that iNexBot's products and technology can improve the efficiency of China's robot industry and empower it."
Currently iNexBot has only two products: the NexDroid platform and the NRC series controllers.
NexDroid is positioned as a robot motion control secondary development platform, providing all the kinematics, dynamics open interfaces and communication protocols a robot needs, supporting customer secondary development to implement custom processes. "NexDroid is the most open platform on the market, supporting both native C/C++ and scripting programming, so customers can develop robots without understanding robot algorithms," Zhang Xiaolong said.
In addition, NexDroid has strong compatibility and adaptability, enabling interconnection with the vast majority of robot components on the market. iNexBot creatively opens up its controller algorithms through NexOpenAPI for other controller or robot body manufacturers to do secondary development — a model already in mass production at multiple customers.
The NRC control system provides a TurnKey solution of controller and teach pendant, with built-in standard applications such as welding, palletizing, cutting and loading/unloading, which can be seen as native applications. The NRC motion controller is capable of controlling almost every robot configuration on the market, supporting dynamics, vision and force control algorithms.
Zhang Xiaolong said that iNexBot's robot controller technology is already leading domestically, but there is still a big gap versus imported products. "We still need 3-5 years to take the control algorithms to the extreme. Next, iNexBot will further increase market share and expand the user base."
Looking at CoDeSys, the benchmark iNexBot aims at, CoDeSys provides the IEC 61131 standard development environment IDE, supports multiple hardware and software platforms (X86/ARM, Linux/Windows), and has very deep technical accumulation.
Zhang Xiaolong also frankly admits that iNexBot still lags far behind CoDeSys in system compatibility. For example, CoDeSys supports the PLCOpen standard, but iNexBot hasn't completed that yet; CoDeSys has also invested heavily in Safety. NexDroid needs to make more efforts in platform compatibility and standardization to fully replace CoDeSys.
Beyond the gaps, Zhang Xiaolong also clearly sees iNexBot's advantages: system usability and the convenience of secondary development. "We're closer to a turnkey solution, with all the kinematics and dynamics algorithms robot industry customers need — especially process application algorithms, where CoDeSys is quite weak. Therefore, CoDeSys places high demands on customers' development capabilities, and most companies cannot quickly digest CoDeSys's technical potential," Zhang Xiaolong further noted.
To close the gap, at the strategic level iNexBot will uphold the principle of "customer-centric", enabling customers at every stage — regardless of R&D strength or company size — to carry out customization at a certain level.
At the technical level, iNexBot will continue to increase technology investment and R&D intensity, expand the R&D team, and narrow the technology gap with international companies. In addition, iNexBot has joined international standards organizations such as EtherCAT and CLPA, applied for the Ministry of Science and Technology's major R&D program for intelligent robots, and is actively participating in the drafting of national robot standards.
Building an Ecosystem
"With iNexBot's control system at the core, we have built an ecosystem that includes component suppliers, body manufacturers, integrators and end users, greatly reducing R&D and business costs for upstream and downstream companies and enabling seamless connection between body manufacturers and integrators," Zhang Xiaolong said.
Reportedly, the process richness of iNexBot's control platform even surpasses that of many imported robots, its open interfaces are more comprehensive and complete than those of domestic competitors, and it already has multiple concrete application cases. iNexBot products are already running on the production lines of large manufacturers such as Huawei.
In product design, thanks to the team's smartphone industry DNA, iNexBot's design philosophy is distinctive: a fully guided operation interface with extremely low training costs — usable right out of the box — which greatly reduces the cost of promoting and applying robots.
In kinematics algorithms, iNexBot has many unique algorithms for smooth trajectory transitions and short line segments — for example, supporting both NURBS and Catmull spline interpolation to balance trajectory fidelity and smoothness. This feature is very important for industries with extremely high trajectory accuracy requirements, such as cutting and adhesive dispensing.
In dynamics, iNexBot has an independently developed load identification method that accurately builds robot dynamic models for collision avoidance and dynamic feedforward, improving operational safety and greatly improving trajectory accuracy. In addition, for heavy-load robot control, iNexBot uses a feedforward control algorithm based on dynamic models to suppress the impact of load changes on the robot, combined with feedback linearization to achieve linearized, decoupled control of heavy-load robot dynamic models, and uses robust control algorithms combined with active disturbance rejection to suppress the inaccuracy of robot model parameters — greatly improving the high-speed dynamic response of heavy-load robots.
Looking ahead, to address the problems of insufficient intelligence in domestic robots, the lack of heavy-load robots, and how robots should be applied in the 5G era, iNexBot will develop fully closed-loop robot control based on vision and force feedback, optimized dynamic control algorithms for heavy-load robots, and an industrial internet based on 5G and TSN technology — further enhancing the market competitiveness of domestic robots.
From the perspective of task completion, robot control is currently still open-loop at the tool end — like a puppet without senses or a brain — so application scenarios are extremely limited. But with vision and touch (force control), robot demand will be thoroughly activated. "iNexBot already has built-in vision and force control algorithms. In the future, we will further integrate deep learning and other algorithms to improve robots' autonomous decision-making, explore new application scenarios, and make robots smarter."
Compared with imported robots, domestic robots have one obvious disadvantage: a lack of heavy-load robots and slow production cycle times. In Zhang Xiaolong's view, this problem stems from poor body consistency and missing dynamics algorithms: compared with light-load robots, non-linear coupling and inertia torque account for a larger share in the dynamic models of heavy-load robots, and traditional PID control cannot achieve high-speed dynamic response.
As mentioned above, iNexBot has already conducted certain R&D on dynamics algorithms for heavy-load robots, and next it will further optimize the control algorithms for heavy-load robots.
In the fieldbus era of industrial control, networking technology for automation products such as robots and PLCs was monopolized by automation giants like ABB, Siemens and Rockwell. Now, China has fully taken the lead over established developed countries in communication technology, especially 5G. As the TSN protocol championed by Huawei occupies more and more industrial application scenarios, iNexBot — as a representative domestic robot controller manufacturer — will fully embrace 5G and TSN technology, making full use of the high real-time performance and big data they bring to lead the application of Chinese technology and standards in the robotics field.
Is the platform model right for iNexBot at this stage?
In the past two years, the ecosystem has been frequently mentioned in the industrial robot industry, but industry insiders generally believe that building an ecosystem is something for listed companies or big enterprises with financial strength. For a startup like iNexBot, is this "aiming too high" or "far-sighted"?
"We're still talking about survival — we have no right to talk about vision." That's how some companies joke. Will iNexBot, which looks up and sees the moon, neglect the sixpence at its feet?
Some industry insiders point out that promoting an operating system in the industrial field is far harder than in smartphones. Industrial scenarios are extremely fragmented, making it hard for tech startups to form true competition with each other. At least in the intelligent robot field, there are rarely large enough scenarios to attract companies to swarm in.
Zhang Xiaolong himself knows that building an industrial robot control system is quite different from building a phone operating system. For B2C companies, practical value may not be so important — once a product has an attractive trait, the impulse to buy follows. But B2B is completely different: you must build long-term value relationships with customers, and product deployment is extremely difficult. "The most important difference is that B2B products must help customers make money, solve real production problems, and deliver a sense of value."
What a platform needs is precisely a customer base — without large-scale deployment, everything is futile. But this process is undoubtedly long, and before that, iNexBot still has to solve the problem of how to survive.
Reprinted from GG Robot (Gao Gong Robot).
