天眼查App显示，12月4日，苏州康阳机器人有限公司成立，法定代表人为陆亚洲，注册资本1亿人民 币，经营范围包括智能机器人的研发、智能机器人销售、服务消费机器人制造、电机及其控制系统研发 等。股东信息显示，该公司由华之杰全资持股。 ...

11月14日，任正非在位于上海的华为练秋湖研发中心与ICPC全球优胜者及教练座谈，围绕AI未来、教 育本质与青年成长展开对话。他强调"教育是教育，商业是商业"，并指出人工智能应聚焦未来三至五年 的产业应用，推动工业、医疗等领域的实质性进步。面对时代浪潮，他鼓励青年"敢于在质疑中前进"， 在开放合作中推动文明交融与技术共进。 任正非在对话中提到，最近小鹏机器人出来走"猫步"，大家 开始不相信，说里面是个真人。小鹏在发布会上当场拿剪刀剪开机器人的外皮，里面都是"钢铁"。机器 人的产业也发生了很大的进步，这些小公司都很厉害，代表了未来的世界，展示一个未来世界。做真正 的"人"还很难，脑的能耗比、皮肤的神经密度……。这个世界总要有人向未来探索，人类社会就是从一 次又一次的失败中成功起来的。中国三、五年后会有较大的进步。中国强大了，有利于世界繁荣的。 ...

新一轮科技革命浪潮下，中国崛起了一批科技新秀，"杭州六小龙"更是集体出圈。宇树科技创始人王兴 兴说："最近一年，具身智能领域有点像做梦的感觉，很多科幻中的故事已变成现实。这种现象在未来 几年还会加速。" 数据是信息时代的宝贵资产，如何用好数据、更好实现数字化是很多小微企业的困扰。针对这一问题， 位于上海的黑湖科技专门设计"黑湖小工单"，抓住"工单"这一企业核心数据，打造了无需复杂配置即 可"分钟级上线"的生产管理小程序。某五金加工厂使用后，订单交付准时率从65%提升到92%，生产效 率提升35%，库存周转率提高40%。 中新社北京12月2日电 (记者 王梦瑶)2024年中国共有27座城市地区生产总值超过1万亿元(人民币，下 同)，被称为"万亿之城"；其中，长三角地区占到三分之一。根据温州近日发布信息，今年该市经济总 量预计可破1万亿元。若实现，长三角地区"万亿之城"有望达到10座。 温州这一历史性跨越，是长三角地区众多依靠民营经济"起家、发家、立家"城市的缩影，如此"群星闪 耀"的长三角地区托起了中国民营经济的关键一极。 从今年前三季度经济数据来看，上海规模以上民营企业工业总产值同比增长9.8%，增速快于全市 ...

Core Insights - The article discusses the limitations of existing robotic manipulation models that primarily rely on 2D images, which often lose critical depth information and 3D geometric structure [2][4] - The proposed solution, SpatialActor, focuses on "disentanglement," separating semantic information from spatial geometric information to enhance robotic understanding and interaction with 3D environments [4][7] Methodology and Architecture - SpatialActor employs a dual-stream architecture that decouples visual and depth encoding, integrating a Semantic-Guided Geometry Module (SGM) and a Spatial Transformer (SPT) to improve robustness and accuracy in robotic tasks [10][11] - The SGM combines robust geometric priors from a pre-trained depth estimation model with fine-grained but noisy depth features, optimizing the geometric representation while maintaining alignment with semantic cues [11][13] - The SPT establishes precise 2D to 3D mappings and integrates multi-modal features, crucial for generating accurate robotic actions [13] Experimental Results - SpatialActor achieved an average success rate of 87.4% across various tasks in simulation, outperforming the previous state-of-the-art model RVT-2 by 6.0% [16][19] - In noise experiments, SpatialActor demonstrated superior robustness, with average success rates improving by 13.9%, 16.9%, and 19.4% under light, medium, and heavy noise conditions, respectively [19][20] - Real-world experiments showed SpatialActor consistently outperforming RVT-2 by approximately 20% across various tasks, confirming its effectiveness in complex environments [22][24] Conclusion - The article concludes that SpatialActor represents a significant advancement in robotic manipulation by effectively decoupling semantic and geometric information, leading to improved robustness and generalization in diverse conditions [24][25] - The framework highlights the importance of disentangled spatial representations for developing more resilient and adaptable robotic systems [25][26]

"这次参会不仅为我们打开了技术视野，更坚定了公司'立足农业实际场景'的创新理念。"12月4 日，"智汇具身 创领未来：具身智能技术创新与场景应用对接会"（以下简称"对接会"）在北京举办。 会场外，北京中科原动力科技有限公司品牌经理张光辉向嘉宾们介绍着农业机器人模型，话语间满是自 豪。 工业制造、现代农业、养老健康……本次对接会不仅发布了具身智能应用的优秀案例，更吸引十几 家企业携机器人亮相。 "我们希望以对接会为'桥梁纽带'，促成更多务实的合作，共同推动我国具身智能技术实现跨越式 发展。"对接会主办单位科技部新质生产力促进中心主任邢怀滨表示。 前沿技术引领产业变革 具身智能是人工智能（AI）的重要落地方向之一。形象地说，它除了拥有聪明的"大脑"，还拥有真 实可感的"四肢"，能够将感知、行动与认知深度融合。各类人形机器人就是具身智能的典型代表。 "具身智能不仅是技术赛道，更是关乎未来智能化产业格局的变革，是推动新质生产力发展的重要 引擎。"对接会主办单位场景创新杂志社主要负责人说，近年来，具身智能领域创新成果正加速涌现。 "超声与X光、CT不一样，它是手艺活+体力活。目前，我们针对心血管超声，创新出全球首个完 ...

12月4日，智通财经记者获悉，大晓机器人将于12月18日正式对外亮相，并开源"开悟"世界模型3.0 （Kairos 3.0）。大晓机器人称，该模型是首个国内开源且已实现且实现商业应用的世界模型。同时， 其将同步发布具身超级大脑模组A1，搭载纯视觉无图端到端VLA具身智能模型，具备自主导航能力。 智通财经记者获悉，商汤科技联合创始人、执行董事王晓刚将出任大晓机器人董事长，世界级AI科学 家陶大程担任大晓机器人首席科学家。 ...

Group 1 - The article emphasizes the investment value of significant events, analyzing industry chain companies and interpreting key points of major policies [1] - The company is involved in humanoid robots and AI glasses, launching multiple products that penetrate the domestic leading robotics market, and investing in the world's first company to mass-produce 1,000 high-freedom dexterous hands, which fill a domestic gap with core products compatible with AI glasses [1] - The company is also engaged in space computing and commercial aerospace, deeply binding with domestic leading aerospace enterprises to provide satellite computing power and AI application solutions, with main products targeting the digital transformation of industrial enterprises [1]

Core Insights - The robotics sector is experiencing a rebound after a significant pullback, driven by multiple marginal benefits and increased confidence from major tech companies like Google and Tesla accelerating their investments in the industry [1][2] - A complete industrial ecosystem is forming, encompassing manufacturing, core component supply, and application scenarios as the supply chain matures [1] Industry Developments - Tesla plans to establish a dedicated facility in Texas with an annual production capacity of 10 million units by 2027 [1] - Yushun Intelligent's emergency robot industrial park project has completed registration, aiming for a monthly production capacity of 500 units [1] - UBTECH has begun mass deliveries of the Walker S2 robot, targeting an annual production capacity of 5,000 units by 2026 and 10,000 units by 2027 [1] - Domestic companies like Zhiyuan Robotics, Galaxy General, and Xiaopeng Robotics are also expanding their production capabilities [1] Supply Chain Progress - Tesla's supply chain (T-chain) is entering a critical verification phase, with expected orders to suppliers in December and guidance for next year's supply [1] - Domestic supply chain companies are rapidly emerging, achieving technological breakthroughs in products like frameless torque motors and precision gears, and have begun small batch deliveries to leading clients [1] Application Scenarios - Application scenarios such as logistics, home services, and data collection are gradually being implemented, with UBTECH winning multiple projects for humanoid robot data collection and training centers [2] - Changan Automobile plans to launch its first vehicle-mounted robot component in Q1 next year, promoting diversified applications [2] Investment Trends - Global capital interest in humanoid robotics is at an all-time high, with increasing financing scales and participant numbers [2] - Domestic companies like Leju Robotics have completed nearly 1.5 billion yuan in Pre-IPO financing, while international firms like Apptronik secured $331 million in strategic financing [2] - Recent ETF initiatives by institutions like Huaxia Fund and E Fund aim to introduce long-term capital support to the sector [2] Focus Areas for Investment - Emphasis on sectors with confirmed industrial certainty, particularly high-probability suppliers in Tesla's supply chain, such as Sanhua Intelligent Controls and Top Group [2][4] - Attention to incremental segments related to technological upgrades, particularly in dexterous hands, with high growth potential for companies like Mingzhi Electric and Dechang Electric [3] - Identifying undervalued companies with substantial progress in robotics, such as Obit Optical and Dongmu Co., which may have a discrepancy in market perception [4]

Core Viewpoint - The robotics sector is experiencing a rebound after a significant pullback, driven by multiple marginal benefits and accelerated investments from tech giants like Google and Tesla, which boosts industry confidence [2] Industry Developments - Major companies are ramping up production capabilities, with Tesla planning to build a dedicated facility in Texas with an annual capacity of 10 million units by 2027 [2] - Yushun Intelligent's emergency robot industrial park project aims for a monthly production capacity of 500 units [2] - UBTECH has begun mass deliveries of the Walker S2 robot, targeting an annual production capacity of 5,000 units by 2026 and 10,000 units by 2027 [2] - Domestic companies like Zhiyuan Robotics, Galaxy General, and Xiaopeng Robotics are also expanding their production capacities [2] Supply Chain Insights - Tesla's supply chain (T-chain) is entering a critical validation phase, with expected orders to suppliers in December and guidance for next year's supply [2] - Domestic supply chain companies are rapidly emerging, achieving technological breakthroughs in products like frameless torque motors and precision gears, and have begun small batch deliveries to major clients [2] Application Scenarios - Applications in logistics, home services, and data collection are becoming more prevalent, with UBTECH winning multiple projects for humanoid robot data collection and training centers [3] - Changan Automobile plans to launch its first vehicle-mounted robot component in Q1 next year, promoting diversified applications [3] Investment Trends - Global capital interest in humanoid robotics is at an all-time high, with increasing financing scales and participant numbers [3] - Domestic companies like Leju Robotics have completed nearly 1.5 billion yuan in Pre-IPO financing, while international firms like Apptronik secured $331 million in strategic financing [3] - Recent ETF filings by institutions like Huaxia Fund and E Fund aim to introduce long-term capital support to the sector [3] Focus Areas for Investment - Emphasis on companies with verified capabilities and batch supply potential, particularly those in Tesla's supply chain, such as Sanhua Intelligent Controls and Top Group [4] - Attention to incremental segments related to technological upgrades, particularly in dexterous hands, with high growth potential for companies like Mingzhi Electric and Dechang Electric [4] - Consideration of undervalued companies with substantial progress in robotics, such as Obit Light and Dongmu Co., which may have a perception gap in the market [4]

编辑丨 量子位 点击下方 卡片 ，关注" 具身智能之心 "公众号 >> 点击进入→ 具身 智能之心 技术交流群 更多干货，欢迎加入国内首个具身智能全栈学习社区： 具身智能之心知识星球（戳我） ，这里包含所有你想要的! 叶问蹲、跳舞、跑步，一个策略全搞定！ 核心思路： AMS从三个关键方面解决动态运动与平衡控制的统一问题： 1. 异构数据源 ：从机器人动作空间直接采样生成可扩展的平衡数据，突破人类数据限制，缓解长尾分布问题。 近日，来自香港大学、NVIDIA和清华大学的联合研究团队提出了一种名为 AMS （ Agility Meets Stability ） 的统一人形机器人全身控制框 架， 首次 实现了在单一策略中同时具备动态运动跟踪和极限平衡控制能力。 2. 混合奖励机制 ：选择性应用平衡先验奖励，精准平衡指导不牺牲敏捷性，化解优化目标冲突。 3. 自适应学习策略 ：动态调整采样概率，同时对每个动作"因材施教"，实现高效的自适应学习。 下面来看详细内容。 人形机器人的"两难困境" 人形机器人要在人类环境中执行各种任务，需要同时具备两个看似矛盾的能力： 敏捷的动态运动 和 精确的平衡控制 。 反观人类，却能轻 ...