Application scenario logic: Try structured scenarios first, accumulate data and then intelligently improve

A number of humanoid robot companies have told surging news reporters that the target scenes of humanoid robots will mainly focus on high-risk high-intensity scenes such as industrial special industries, commercial scenes and nursing scenes such as pension and childcare.

IC200ALG630 Chen Jianyu, assistant professor at Tsinghua Institute of Interdisciplinary Information and founder of Star Moving Era, said that the reason for choosing industrial scenes is that the human factor can be put aside, and people will bring the greatest uncertainty, and play its ability in a relatively limited, structured and predictable scene.

Ji Chao, chief scientist of iflytek Robotics, told surging news reporters that the application scenario of humanoid robots will first cut into manufacturing scenes such as automobile manufacturing, vehicle assembly, logistics sorting, and repeat some tasks that are complicated or dangerous, and not particularly complex and not requiring high accuracy, and then slowly accumulate data and iterative algorithms. Finally, intelligent upgrading and entering the home.

Walker S humanoid robot “internship” in car factory

It can be seen that domestic humanoid robots have indeed entered the industrial manufacturing and service industry scenes.

IC200ALG630 This year, the “humanoid robot first share” Youbixuan has reached cooperation with Weilai, Dongfeng Liuqi, Beijing and FAW-Volkswagen and other automakers to complete assembly, testing and other work and the exploration of highly intelligent and flexible production lines and automotive super unmanned factories; Fourier Intelligent humanoid robots have also entered the Shanghai branch of China Construction Bank, cooperated to build a commercial bank scene application training base, the robot completed simple dialogue and communication in the lobby and other work, but also began to “practice” in the SAIC General Motors factory, to carry out the installation of high-voltage parts, high-precision operation and other work.

The company said that by deploying humanoid robots in the real car building environment, it can continuously accumulate total factor production data, use AI technology to perceive, learn and iterate, achieve intelligent management closed-loop, and quickly adapt to new task requirements based on the autonomous learning ability of humanoid robots with embodied intelligence, and provide agile deployment in manufacturing scenarios. At the same time, the automotive manufacturing industry is large and highly automated, and the willingness to introduce humanoid robots is relatively strong.

Yang Zhengye told reporters that the humanoid robot has mobility and versatility, which is suitable for application in factories. Humanoid robots are part of embodied intelligence and can be extended to other aspects in the future. “Because the scene is set, the work of the scene is set. For these scenarios, we can make customized hardware and software, try it first in some areas, and then feed the data and technology back into the universal humanoid robot, and eventually become the closest human assistant.”