news

Xiong Youjun, currentGeneral Manager of Beijing Embodied Intelligent Robot Innovation Center, and has been responsible for many Ministry of Science and Technology Innovation Fund projects and National 863 Program projects.

He has been engaged in research on robot-related technologies for nearly 20 years.Xiong Youjun, led the development of a number of star robot products. He also led the Chinese robot delegation to participate in global robot competitions many times as the head of the delegation, winning 8 world championships and 6 runner-ups.

Participated in the creationBeijing Embodied Intelligent Robot Innovation Center (formerly known as Beijing Humanoid Robot Innovation Center, hereinafter referred to as the Innovation Center)It is his "second venture".

The 2024 World Robot Conference was held in Beijing from August 21 to 25.Xiong YoujunHe was interviewed by a reporter from National Business Daily (NBD).

Humanoid robots will become an important supplement to the human workforce in the future

NBD: In which scenarios do you think humanoid robots will have greater advantages? Can you give some examples?

Xiong Youjun: I think humanoid robots, as general and multifunctional robots, have advantages in many areas.

Putting aside the price, first of all, humanoid robots will become an important supplement to the human workforce in the future. In theory, humanoid robots can be used wherever humans are needed in the future. I have seen predictions that by 2030, China's labor force will be greatly reduced. Even though we have already applied a large number of industrial robotic arms and special robots, it still cannot make up for the labor shortage.

Secondly, many jobs are harmful to people and can easily cause work-related injuries. For example, there are a lot of toxic and harmful gases in the chemical and pharmaceutical fields, and many inspection positions require X-ray inspection. I have been to a car factory to see the inspection process. After the car is assembled, workers need to crane their necks to check the assembly points, working 8 hours a day. There are also many assembly points that require thumbs to press, which may cause work-related injuries over time.

In addition, many types of work are dangerous and unsuitable for humans. For example, the petrochemical industry and firecracker production companies have flammable and explosive materials, and there is a large market demand for humanoid robots in these fields. UHV substations and underground coal mines also have certain safety risks, and these jobs can be replaced by humanoid robots.

These areas should be the priority areas for the development of humanoid robots. However, the current technology cannot support the large-scale application of humanoid robots in these areas. Once the technology achieves a breakthrough and gradually matures, I think the volume will increase quickly. Therefore, Musk expects that when the development of humanoid robots enters the mature stage in the future, the ratio of the number of robots to the human population will be at least 1:1, and it is likely to reach 2:1, and the number of robots will exceed the number of people. This is not just talk, there is a whole set of logic behind it, and there is also a lot of preliminary research.

Why are many well-known foreign companies investing in robotics companies, and many domestic companies are also starting to invest? These are not short-sighted decisions, but the result of careful consideration and market research. Companies will consider long-term development.

Of course, we actually talked a lot about ethical issues at the forum today. I think we must also consider the impact of this aspect on society in the future. The advancement of the industry must be people-oriented, and we must never ignore people’s social needs. This will have no vitality.

The development of humanoid robots is still in its early stages and many underlying core issues need to be addressed

NBD: What do you think is the most urgent technological breakthrough needed in humanoid robots today?

Xiong Youjun: Actually, there are a lot of them. I think it is difficult to summarize them in one or two sentences, because the development of humanoid robots is still in a relatively early stage even now.

For example, the motion control technology of humanoid robots. Although we have seen that many humanoid robots can do somersaults and run fast, when they really enter the ever-changing daily life or work scenarios, there are still many stability, reliability, and applicability issues that need to be solved.

Another example is the "brain" of a humanoid robot. Although humanoid robots now have large models and embodied intelligence, they are not smart enough. The robot's ability to interact naturally with humans, its decision-making ability, and its task planning ability will indeed take a long time to develop.

Of course, there are still many underlying core issues that need to be resolved. For example, in terms of embodied intelligence, the construction of data sets is very important. Traditional artificial intelligence technology has borrowed a lot of data from the mobile Internet, with a lot of text and pictures fed to it for learning, which has enabled artificial intelligence technology to advance rapidly.

The same is true for embodied intelligence, which requires a large amount of data. However, this data is no longer just text and pictures, but also requires a large amount of operational data. Because embodied intelligence is more about interacting with the physical world, it needs to get you a glass of water, install, assemble, and carry with you, which involves a lot of operations and requires a lot of behavioral data, which is scarce. Only after this data is built can the embodied intelligence model be well trained and the robot can truly have a smart "brain".

Now the country is guiding innovation centers like ours to lay a solid foundation and enable enterprises to apply and implement them. Through division of labor and cooperation, we can jointly promote the development of the humanoid robot industry.

NBD: With the trend of integration of artificial intelligence and robotics technology, can humanoid robots achieve a higher level of intelligent interaction with the help of technologies such as deep learning and natural language processing?

Xiong Youjun: Of course, this is necessary, because the development of embodied intelligence is inseparable from the progress of traditional artificial intelligence technology, especially in human-computer interaction and task planning, traditional artificial intelligence has great advantages.

Daily Economic News