人工智能具身智能数据生成平台
Search documents
全国首个!具身智能领域“上海虚实融合具身智能训练场”国家级标准化试点落沪
机器人大讲堂· 2025-09-30 04:47
Core Viewpoint - The establishment of the "Shanghai Virtual-Real Fusion Embodied Intelligence Training Ground Standardization Pilot" marks a significant advancement in the field of embodied intelligence in artificial intelligence, aiming to unify training standards across the nation and promote data collection and ecological collaboration among various provinces and cities [1]. Group 1: National Standardization Pilot - The pilot is led by the Human Robot (Shanghai) Co., Ltd., which will oversee the main training ground in Shanghai and support various sub-training grounds across the country, implementing over 80 standards [1][5]. - The pilot aims to create a "1+N" model, where the main training ground serves as the standard hub, facilitating the implementation and iterative optimization of standards across sub-training grounds [6][14]. Group 2: Standardization Framework - The pilot will establish a "basic + characteristic" standard system focusing on four key areas: physical entities, virtual training grounds, components, and application scenarios [6]. - The initiative will involve the development of national standards for humanoid robot datasets and the integration of these standards into international standards [8]. Group 3: Industry Collaboration - The pilot will collaborate with various innovation centers across provinces to develop industry standards, including training data management and environmental construction specifications for humanoid robots [12]. - The project aims to create a closed-loop process for training, standards, testing, and certification, enhancing the quality and reliability of data collected from training grounds [8][12]. Group 4: Data Collection and Utilization - The pilot anticipates collecting 150,000 hours of high-quality real machine data by 2025, with an annual production capacity exceeding one million hours [14]. - Various sub-training grounds will focus on specific regional industries, such as textiles, automotive parts, and renewable energy, to gather high-quality data tailored to local needs [14]. Group 5: Future Outlook - The pilot is seen as a critical turning point for the embodied intelligence industry in China, transitioning from isolated innovations to a collaborative ecosystem [17]. - The goal is to establish a unified innovation force by integrating disparate industry strengths, positioning China as a leader in the global embodied intelligence sector [17].