The First Heavy-duty Robot Forum of Blue-Orange Laboratory: There are still three major challenges in the intelligentization of industrial robots, and they won't be completely replaced by humanoid robots.

Author | Huang Nan

Editor | Yuan Silai

In the wave of the large model era, the industrial robot field is undergoing another round of reflection and transformation.

Recently, the first Academic Forum on Heavy - Duty Robots hosted by the national science and technology innovation platform, the Blue - Orange Laboratory, was held in Foshan. Many industry experts conducted in - depth discussions on topics such as the application of industrial robot technology and equipment, exploration of intelligence, and integration with large models.

The Blue - Orange Laboratory is a national key laboratory for heavy - duty robots established by Midea Group in 2022, focusing on key areas such as core components of robots, overall machine design, intelligent control, and intelligent applications.

High - level Dialogue

As a major direction in the industrial robot field, with the transformation of high - end manufacturing to robotic equipment, heavy - duty robots have become key flexible intelligent industrial mother machines in multiple industries, and market demand has soared. Technology giants represented by Midea have continuously increased their layout in this field in recent years and have also made some progress.

As a flexible multi - body system, heavy - duty robotic arms need to achieve higher repeat positioning accuracy under heavy - load and high - speed start - stop conditions. In some scenarios, absolute positioning accuracy and trajectory accuracy are also required.

Chen Wenjie, the executive deputy director of Midea's Blue - Orange Laboratory, pointed out that currently, heavy - duty robots for processing scenarios need to achieve a leap - forward increase in stiffness and accuracy. That is, the stiffness of the end of the whole machine needs to be increased by 3 to 5 times, and the absolute accuracy with load also needs to be doubled synchronously.

To achieve this goal, based on the V - shaped model development process, the Blue - Orange Laboratory optimizes according to the needs of application scenarios through the forward - design simulation platform, and proposes a topological concept of heavy - duty robots and a design of the transmission chain that are conducive to a leap - forward increase in accuracy and stiffness.

Since acquiring KUKA in 2017, Midea has focused on solving the "bottleneck" problems of key components of heavy - duty industrial robots and promoting integration in vertical fields. Currently, the market share of Midea KUKA's heavy - duty robots has reached 41.5%, holding an absolute market advantage in the industry.

In addition, Midea has also laid out collaborative robots, such as seven - axis collaborative robots. Their accuracy and collaboration ability are stronger than previous models, and they have been widely used in multiple fields such as industry, medical care, and consumer electronics.

According to official data, in 2024, the business revenue of Midea's robotics and automation segment exceeded 30 billion yuan. It owns brands such as KUKA, Swisslog, Servotronix, and Jiyajingji, achieving full coverage from core components to overall solutions, and from hardware equipment to intelligent systems. It has become one of the top four industrial robot manufacturers in the world.

Nowadays, most people regard industrial robots as a part of traditional industries. In fact, with the integration of AI technology and hardware, within the existing framework system, large language models can replace manual programming. The flexibility, programmability, and intelligence level of industrial robots have been improved, as well as production efficiency and quality.

"Just like autonomous driving, with the help of embodied intelligence and large models, robots can have the ability of autonomous operation and learning, and complete manufacturing processes and operation tasks in the industrial manufacturing environment," said Zhang Hui, the Chief Technology Officer of Midea KUKA Group.

However, industrial robots also have their limitations. Currently, they still tend to be industrial equipment and have a large gap from human intelligence. Chen Heping, a professor at the University of Hong Kong, pointed out that "robots in the new era face three major challenges: intelligent programming, intelligent challenges, and human - robot collaboration. Moreover, robots currently have few closed - loops with processes, while humans can continuously improve their skills through closed - loops."

Chen Heping believes that in the future development of industrial robots, their performance in terms of intelligence needs to be improved. At the same time, although humanoid robots have attracted much attention and huge investment, their development is closely related to application scenarios. Different forms of robots, such as dual - arm robots, bipedal robots, and mobile robots, all have their own application scenarios, and industrial robots still have an irreplaceable position and will not be completely replaced by humanoid robots.

Zheng Feng, the deputy dean of the Robotics Research Institute of Southern University of Science and Technology, also said that although embodied intelligence is developing rapidly, in complex, high - precision, and high - intensity scenarios such as industrial manufacturing, traditional industrial robots have more advantages with algorithms driven by professional knowledge, because it is difficult to obtain large - scale data to train models in these scenarios.

According to the "World Robotics Report 2024" of the IFR, in 2023, the installation volume of industrial robots in China reached 276,000 units, accounting for 51% of the global installation volume; the stock of industrial robots in operation in China was close to 1.8 million units. China is the first and only country in the world with such a large stock of industrial robots.

Industrial robots are born to solve specific problems, and the methods and tools for solving problems are also constantly being updated. In the process of development from machines to humans, the update of methods prompts them to continuously solve new problems, and new forms also emerge accordingly.

In the era of large models, exploring new development paths, whether it is humanoid or not is not the key. The important thing is that industrial robots are always moving forward, bringing more possibilities to the industrial field.