|
listen to this article  |
China has launched a data sharing initiative for domestic humanoid companies to accelerate innovation. This project is being led by the National Joint Local Humanoid Robot Innovation Center and various humanoid companies in China.
As humanoids mature, they will require vast amounts of training data to enable their intelligence to execute routine tasks autonomously. Every developer has the same problem. How do you generate training data for these systems? China recognizes this problem and is seeking ways to overcome it by bringing together its robotics industry.
Announcement of Chinese guidelines
The National Local Joint Humanoid Robot Innovation Center (NLJIC), China’s first public platform specializing in this type of robot, is establishing a training ground for manufacturers to collect high-quality data and promote industry standards. Xu Bin, general manager of the National Support Center, told Chinese media outlet Yicai Global that like self-driving cars, “humanoid robots also require large amounts of high-quality data for development, and advanced models require more stringent data standards.”
China has already published two standard documents for humanoids.
-
- Guidelines for classification and application of humanoid robots
- Guidelines for Categorizing Embodied Intelligence Smart Development Stages
At the time of reporting, robot report Still trying to find a copy of this document.
A team led by NLJIC has published an industry standard for humanoids in China. The standard classifies robots into four skill levels (L1 to L4) and five intelligence levels (G1 to G5), assessing capabilities such as perception, cognition, decision-making and autonomy.
NLJIC was established in Shanghai in May 2024. The center also announced its first full-size humanoid robot, Qinglong, at the 2024 World Conference on Artificial Intelligence (WAIC). Qinglong is designed as an open source platform to advance the humanoid field globally.
Apply to speak.
NLJIC aims to train 1,000 humanoids by 2027. Several Shanghai-based companies, including Fourier Intelligence and Kepler, are developing humanoids that would benefit from the data-sharing initiative. Fourier Intelligence launched the second generation GR-2 in September 2024.
IEEE Humanoid Research Group
A similar initiative is underway under the mentorship of Aaron Prather, ASTM International’s robotics and autonomous systems program director. The IEEE Humanoids Research Group is a year-long study whose goal is to evaluate all current robotics safety standards, determine which standards cover humanoids, and identify gaps for future SDOs to begin the process of creating new, applicable standards. This is an ongoing project.
“China has clearly made humanoids a national priority,” Prather said. “This should be a wake-up call to other countries. “It also justifies the work of the IEEE Humanoids Research Group and its work to develop humanoid standards at an international level to help accelerate humanoid adoption everywhere, not just in single countries.”
