According to statistics, China has been the world's largest robot application market for five consecutive years, but high-end robots still rely on imports. The gap between the core algorithms is a roadblock for domestic industrial robots to move toward high-end manufacturing.
The algorithm has a lot of debts. The domestic "brain" loves mistakes.
As an industrial grade product, there are two main criteria for measuring the pros and cons of a robot: stability and accuracy. The core controller is a key component that affects stability and is known as the “brain” of industrial robots. The software is equivalent to language, and the idea of "brain" is passed out.
The gap between the core algorithms is too large, resulting in poor stability of the domestic robots and high failure rates. “Parameter self-tuning, vibration suppression algorithm, torque fluctuation compensation...” Sun Wei (a pseudonym), a technical director of a domestic robot enterprise, glared at his finger, and an item was given to the Science and Technology Daily reporter. “The gap in the algorithm has slowed down the entire system. Response speed and stability."
Therefore, some customers prefer to buy second-hand imported robots, pay a large amount of "reboot fee", and are not willing to use domestic robots.
"A good algorithm, thousands of lines can make the robot run stably without failure; the poor algorithm, tens of thousands of lines can not reach the level of others," Zhou Chao said, "just like two drivers driving, just taking a driver's license Novices and experienced racers can see the gap after the road."
Knowing it, I don’t know why, it’s the result of years of owed by the core algorithm.
Without mastering the core algorithm, products with low production precision requirements are still barely available, but if applied to high-end fields such as aerospace and military, they can only rely on imported industrial robots.
Software card neck pricing power
The gap between the algorithms is not only reflected in the core controller, but also slows down the response speed of the servo system.
Each time the robot completes an action, it needs the core controller, servo drive and servo motor to cooperate. For example, a "war", "general" ordered an attack; "transfer" transmitted an offensive signal; the "soldier" ordered to charge to the designated position.
Now the products of the "Four Big Family" have evolved into "Generals" who directly direct "soldiers" through 4G signals; while domestic robots are still in the era of commanders, the speed is of course much slower.
"If you don't master the core algorithm, this gap is hard to narrow down." Malone told the Science and Technology Daily reporter. According to him, for a single servo system, the dynamic and static accuracy of domestic robots are very high, but high-end robots generally have more than 6 servo systems at the same time, and it is difficult to obtain good control effects by traditional control methods.
Multiple servo systems require the "brain" to calculate in advance. Through the underlying algorithm, the foreign core controller can directly operate the motor through the current loop of the servo system to achieve precise control under high dynamic multi-axis nonlinear conditions. Therefore, the "four major families" robots have faster response speed and more accurate positioning.
Due to the gap between algorithms and software, the price of the final product is also different. Taking the servo system as an example, even if the core components are almost identical, the cost is actually almost the same, but the price of foreign products is 10 times more expensive.
"This is because the domestic servo system basically uses its own software library. Foreign companies limit many advanced functions and do not open to us. For example, the S-curve function of the position ring, if you want this function, you have to pay for the products of others. The pricing power has just let it go," Malone said.
Encourage enterprises to invest in research and development
Although the gap is obvious compared with the “four big families”, experts and practitioners are more optimistic about the future of domestic robots.
From a technical point of view, the development of hardware has been for decades, and there are few secrets. In terms of software, China has a large population of IT practitioners. “Now, many of our companies can also integrate high-level industrial robot products. Zhou Chao said, "But domestically produced industrial robots with completely independent intellectual property rights and core technologies of key components still need experience, time and accumulation and polishing of the entire industrial chain."
The industrial robot market is a bit like the smartphone market a few years ago. As long as you are familiar with the industry chain and supply chain, you will soon be able to "smash" a robot product. But to master the core technology is not that easy. When a certain technology makes a breakthrough, waiting for the market to test and slowly mature, foreign manufacturers are likely to lower the price of the parts and provide a more mature and cheaper solution, which leads to the domestic enterprises' investment and return are not proportional. It is also the reason why many companies are reluctant to invest in the underlying research and development.
"As a practitioner, we hope to have more technology-oriented policies. As long as the technology meets the standards, it will give the company a certain market space. In practice, we will slowly polish the products, thoroughly understand the core technology, and encourage enterprises to put more manpower. Financial resources are invested in research and development, so that domestic industrial robots can truly play a leading role in the world stage," Malone said.