With the accelerated deployment of the artificial intelligence industry, how to enable robots to better understand and adapt to complex real-world environments has become a focal point of industry attention. On June 25th, at the Tianfu Artificial Intelligence Industry Ecosystem and Product Launch Conference held in Chengdu, Zhiyuan Innovation (Chengdu) Technology Co., Ltd. officially released its new "Embodied Intelligence Data Collection 2.0" technology system, providing a core driving force for robot development.

Compared to the 1.0 version, the newly released technology system has made significant breakthroughs in three dimensions: data operations, model evaluation, and scenario implementation. Most notably, it introduces a pioneering "real machine + no body" dual-track data collection mode, which greatly improves the efficiency and flexibility of data collection, effectively solving the pain points of high training costs and insufficient sample coverage in traditional methods. In addition, the system has built the first standardized integrated embodied intelligence model evaluation platform in China, establishing an objective "examination system" for the "brain" of robots.

The release of this technology also fills an important industrial gap in the field of embodied intelligence in Chengdu and the southwest region. According to reports, the Zhiyuan Southwest Embodied Intelligence Industrial Base has already been put into operation in Pidu District this May, specifically used for data collection and training. The implementation of this 2.0 version technology will further accelerate the large-scale operation of this industrial base, promoting the embodied intelligence industry in Chengdu and across the country into a new stage of more systematic and refined development.

As embodied intelligence gradually moves from laboratories to various industrial and living scenarios, this technological upgrade by Zhiyuan Innovation not only demonstrates the depth of exploration by domestic AI vendors in the field of robot interaction, but also provides a solid foundational technical support for future development of more high-performance and intelligent robot products.