Home Bots & Brains China Wants More Data-Sharing for Humanoid Robots

China Wants More Data-Sharing for Humanoid Robots

by Marco van der Hoeven

China is advancing data-sharing efforts within its humanoid robotics sector, aiming to streamline development and standardization. This initiative, led by the National Local Joint Humanoid Robot Innovation Center (NLJIC), marks China’s first public platform dedicated to humanoid robots. At an industry forum on Oct. 28, the NLJIC signed a letter of intent alongside companies like Shanghai-based Kupasai Technology, establishing a smart data ecosystem to enhance industry cooperation and data accessibility.

According to Xu Bin, general manager of the NLJIC, humanoid robots, akin to autonomous vehicles, rely heavily on vast quantities of high-quality data for development, with advanced models requiring even more rigorous data quality standards. To address this, the NLJIC plans to set up a training ground to gather high-quality data and encourage industry-wide data standards. Since its founding in May, the NLJIC has been working on a training center in Shanghai’s Pudong New Area, with facilities that can currently support up to 100 robots, projected to expand to 1,000 by 2027.

Shanghai’s Municipal Commission of Economy and Informatization further announced a strategy to bolster the humanoid robotics industry. This includes creating physical and virtual training centers, strengthening the local supply chain, and building a comprehensive network of related businesses to support all stages of the development process.

At the forum, the NLJIC introduced China’s inaugural industry standards for humanoid robots, categorizing these machines across four technical levels (L1 to L4) and five intelligence levels (G1 to G5), based on key functionalities such as perception, cognition, decision-making, and autonomy. These new classifications are expected to provide a structured framework for evaluating and guiding progress in humanoid robotics.

Misschien vind je deze berichten ook interessant