en.Wedoany.com Reported - On March 29, China's Ministry of Industry and Information Technology approved and released the "YD/T 6770—2026 Artificial Intelligence Key Technology Embodied AI Benchmarking Methods." This is the first industry standard introduced by China in the field of Embodied AI and will be officially implemented from June 1, 2026. This standard fills the gap in China's domestic Embodied AI evaluation system, marking the industry's transition from a lack of unified standards to a new stage of standardized and regulated development.
As a cutting-edge field combining artificial intelligence and robotics, Embodied AI is accelerating its penetration from laboratory research into practical scenarios such as industrial assembly, warehousing and logistics, household services, and medical care. The year 2026 is regarded by the industry as the beginning of its large-scale application. However, previous issues such as the lack of an evaluation system, fragmented technical pathways, and difficulties in quantifying product performance led to inconsistent standards among enterprises, incompatible data formats and interfaces, resulting in wasted R&D resources and hindering the industrialization process of technological achievements.
Against this backdrop, the introduction of the new standard holds significant practical importance. Led by the Artificial Intelligence Standardization Technical Committee of the Ministry of Industry and Information Technology and spearheaded by units including the China Academy of Information and Communications Technology, this standard integrates the practical experience from industry, academia, research, and application, establishing a unified, scientific, and operational benchmarking framework.
The standard regulates the testing processes for two major scenarios: simulation environments and real-world environments, clarifying four core components: environment setup, task library construction, testing process, and metric calculation. Specifically, simulation tests must replicate the complexity of real-world scenarios, while real-world tests must account for application diversity; it proposes methods for task library construction and generalization evaluation benchmarks; for models and complete systems, it defines various methods such as static simulation, dynamic simulation, and real-world environment testing; and it establishes five core metrics: task execution efficiency, task success rate, human intervention rate, scene disturbance attenuation rate, and average task energy consumption, providing a basis for quantifying the full-chain capabilities of Embodied AI systems.
An expert from the China Academy of Information and Communications Technology stated: "A credible evaluation system is the bridge and benchmark connecting technological research with large-scale industrial application." The implementation of this standard will provide a unified measurement basis at the technological R&D level, guiding the direction of iteration; support product selection and verification at the application deployment level; and accelerate the transition of technology from the laboratory to real-world scenarios at the industrial development level, promoting engineering transformation.
The influence of this standard has extended to the international arena. It is currently being advanced simultaneously for international standard project initiation within the International Telecommunication Union, echoing the IEEE P3927 series of international standards for industrial Embodied AI led by the China Academy of Information and Communications Technology, jointly promoting the coordinated development of the global evaluation system. This demonstrates China's technological accumulation and industrial advantages in the field of Embodied AI and lays the foundation for participating in global standard-setting.
With the implementation of this standard and the advancement of the "Humanoid Robot and Embodied AI Standard System (2026 Edition)," China's Embodied AI industry is expected to gradually overcome the issue of standardization lag and move towards high-quality development. In the future, the improvement of the standard system and the deepening of collaboration among industry, academia, research, and application will promote the large-scale application of Embodied AI technology in more scenarios, shifting from demonstrative functions to practical value, and providing new momentum for manufacturing upgrades and the optimization of public services.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
