温馨提示 ： 点击下方图片，查看运营团队最新原创报告（共235页） 说明： 欢迎约稿、刊例合作、行业交流 ， 行业交流记得先加入 "机器人头条"知识星球 ，后添加（ 微信号：lietou100w ）微 信； 若有侵权、改稿请联系编辑运营（微信：li_sir_2020）； 正文： 2025 年 8 月 9 日，由 北京人形机器人创新中心 （国地共建具身智能机器人创新中心） 主办的 " 具身智能技术 与应用论坛 " 在 2025 世界机器人大会（ WRC ）期间圆满举行。论坛现场，来自全国近三百名具身智能领域专 家学者与企业代表齐聚一堂，共同见证具身智能从技术突破走向产业落地的重要进程。 在具身智能加速演进的当下，全球机器人产业正迈入从 " 能动 " 走向 " 能用 " 的关键窗口期。随着基础模型能力 跃迁、感知决策系统趋于成熟，以及工业、物流等应用场景需求爆发，技术突破与产业落地的耦合速度前所未 有。 正是在这样的背景下， 北京人形围绕从 " 最能跑 " 到 " 最好用 " 全面进化的目标，聚焦具身智能核心技术 —— 持续攻坚 ，并在本次论坛上首次对外发布。 " 最好用 " 的人形机器人，不仅要具备稳定可靠的行 ...

Group 1 - Beijing is promoting the industrialization of embodied intelligence with new initiatives during the 2025 World Robot Conference, including the release of a social experiment plan and a set of supportive policies for innovation in embodied intelligent robots [1] - The new policies include annual financial support of up to 30 million yuan for enterprises, focusing on key areas such as technology collaboration, data element trials, application scenario promotion, and nurturing new business models [1] - The Beijing Human-Robot Innovation Center announced four core achievements in embodied intelligence, including a comprehensive model system and a real-world data collection plan for robots, aiming to transition from technological breakthroughs to practical industrial applications [1] Group 2 - The revenue of China's robot industry increased by 27.8% year-on-year in the first half of this year, with industrial and service robot production growing by 35.6% and 25.5%, respectively [2] - Beijing's robot industry is among the top in the country, with the humanoid robot sector accounting for about one-third of the national market, and a nearly 40% revenue growth in the first half of the year [2] - The focus of humanoid robots is shifting from extreme mobility to practical applications, as demonstrated by various operational scenarios at the conference [2] Group 3 - The 2025 World Robot Conference featured more humanoid robots and innovation centers, including several regional centers across China [3] - The goal is to compress the commercialization timeline of humanoid robots from approximately 25 years to around 10 years by concentrating resources and fostering collaboration [3] - The current innovation centers emphasize the integration of AI and robotics, highlighting the necessity of these centers in advancing humanoid robots towards practical applications [3]

2025 年 8 月 9 日，由 北京人形机器人创新中心（国地共建具身智能机器人创新中心） 主办的 " 具身智能 技术与应用论坛 " 在 2025 世界机器人大会（ WRC ）期间圆满举行。论坛现场，来自全国近三百名具身 智能领域专家学者与企业代表齐聚一堂，共同见证具身智能从技术突破走向产业落地的重要进程。 在具身智能加速演进的当下，全球机器人产业正迈入从 " 能动 " 走向 " 能用 " 的关键窗口期。随着基础模 型能力跃迁、感知决策系统趋于成熟，以及工业、物流等应用场景需求爆发，技术突破与产业落地的耦合 速度前所未有。 正是在这样的背景下，北京人形围绕从 " 最能跑 " 到 " 最好用 " 全面进化的目标，聚焦具 身智能核心技术持续攻坚，并在本次论坛上首次对外发布。 " 最好用 " 的人形机器人，不仅要具备稳定可 靠的行动能力，更要在真实场景中实现感知、决策、控制与迭代优化的全链路协同。这一目标的前提，是 同时突破认知理解、通用适配、执行控制与数据供给等核心环节。为此，北京人形在论坛现场正式发布了 四项具身智能核心成果 ——" 具身世界模型体系 " 打通认知与预判能力， " 跨本体 VLA 模型 " 实现 ...

Core Insights - The Beijing Humanoid Robot Innovation Center unveiled four core achievements in embodied intelligence at the 2025 World Robot Conference, marking a shift from technological breakthroughs to industrial practicality [1][2] - The "Embodied World Model System" serves as the core brain for robots, enabling them to understand reality and predict changes, featuring a 72B multimodal large model and a neural network-driven world simulator [1] - The "Cross-ontology VLA Model" allows for flexible adaptation of robots across multiple ontologies and scenarios, demonstrating robust generalization capabilities and significant potential for rapid skill acquisition [1][2] Industry Challenges - Current VLA models face challenges such as action prediction difficulties, significant differences in robot ontology, data incompatibility, and weak task generalization [2] - The direct mapping of image pixel space to robot operation space presents optimization challenges and prediction inaccuracies, complicating the application of models across different robot configurations [2] Data Collection Initiative - The "Thousand Robot Real Scene Data Collection Plan" aims to enhance core capabilities by deploying robots in real industrial environments, collecting multimodal interaction data during actual operational tasks [2] - This initiative contrasts with laboratory settings, as real-world environments introduce variability and unexpected conditions, providing richer training material for embodied intelligence models [2] Future Outlook - Industry experts agree that industrialization is essential for the development of embodied intelligence, emphasizing the need for a standardized and replicable technological leap to achieve scalable implementation [2]