Core Viewpoint - The current landscape of dexterous hand technology is characterized by diverse approaches from different manufacturers, with no unified solution yet established, reflecting a balance between performance, cost, and reliability [1] Summary by Sections Trends in Dexterous Hand Technology - The core area of differentiation in technology routes is the transmission system, with Tesla's Optimus evolving from a "worm gear + tendon" solution to a rumored "screw + tendon" hybrid transmission for its fourth generation, enhancing freedom while optimizing force transmission efficiency [2] - Domestic companies primarily use linkage transmission, such as the RH56BFX series from Yinshi Robotics, which achieves 11 degrees of freedom through linkage and spring transmission [2] - Some companies have broken through the limitations of previous solutions, like Zhiyuan Robotics' five-finger dexterous hand using a mixed scheme of screw + linkage + turbine rod + tendon, indicating a potential return to the mainstream for the screw + tendon composite scheme [4] Technical Innovations and Challenges - Innovations have addressed previous bottlenecks of tendon solutions, with domestic manufacturers finding better solutions to issues like precision control and wear [6] - The use of ultra-high molecular weight polyethylene fiber (UHMWPE) in tendon materials has shown promise, with Chinese production accounting for nearly 50% of global supply [8] - Tactile sensors have become standard in dexterous hands, with various implementations enhancing environmental interaction capabilities [8][10] - The main technical challenges lie in hardware and software, particularly in achieving high degrees of freedom within compact structures while controlling costs [10][12] Market Dynamics and Player Strategies - The dexterous hand market participants can be categorized into three main groups: self-research entities, new dexterous hand forces, and component extension players, each leveraging unique advantages to occupy niche markets [15] - Self-research entities focus on optimizing overall machine performance through in-house dexterous hand development, while new forces prioritize core technology breakthroughs and rapid market entry [17] - Component extension players leverage supply chain advantages, with companies like Zhaowei Electromechanical and Leisai Intelligent developing dexterous hands that integrate self-researched micro-reducers and sensors [19] Future Outlook - The industry is expected to experience significant growth, driven by continuous investment, policy support, and the expansion of application scenarios [22] - Deep bionic design is seen as a long-term direction for dexterous hand development, aiming to fully simulate human hand functions and reduce interaction costs [22] - The integration of AI models with dexterous hands is anticipated to enhance operational skill learning and adaptability, supporting the scale application of dexterous hands in various fields [24][25]

Group 1 - Zhejiang University developed a high-mobility quadruped robot named "White Rhino," which set a Guinness World Record by completing a 100-meter sprint in 16.33 seconds. The robot's design involved precise dynamic modeling and multi-objective optimization algorithms to enhance its performance [3][4] - The robot features a high-power density joint actuator that provides strong torque and quick response, akin to a "race car" muscle system. Its intelligent movement capabilities are driven by a dynamic control strategy based on reinforcement learning algorithms [3][4] - Future applications for "White Rhino" include disaster rescue and transportation in extreme terrains, expanding its utility beyond just speed [3] Group 2 - NVIDIA introduced a new world AI model called Cosmos Reason, designed for robotics and physical applications, featuring 70 billion parameters. This model enhances the reasoning capabilities of robots and AI agents, allowing them to plan actions based on memory and physical understanding [6][7] - The Cosmos model series also includes Cosmos Transfer-2, which accelerates the generation of synthetic data from 3D simulations, aimed at training robots and AI agents [6][7] Group 3 - A team from Northwestern Polytechnical University and Hong Kong universities developed a multifunctional soft robot based on a new type of electroactive polymer, suitable for complex structures and extreme environments. This new material significantly reduces heating and enhances dielectric and mechanical properties [9][10] - The new polymer exhibits over 50% reduction in heating, a lifespan extension of more than 15 times, and a 1.75 times increase in output force, achieving a strong adsorption force of 30 kPa at a low electric field of 2 V/μm [9][10] Group 4 - A collaboration between Ocean University of China and North Carolina State University resulted in a novel soft robot that integrates driving and sensing capabilities. This "C-shaped bipedal micro soft robot" can move quickly and detect its environment through vibration sensing [12][14] - The robot, measuring only 11 mm in length and weighing 0.071 grams, can travel at a speed of 42.8 body lengths per second and carry loads up to 3.47 times its weight, making it suitable for applications in hazardous environments [12][14] Group 5 - Guangzhou Yingbo Intelligent Technology Co., Ltd. secured several million yuan in financing for its service robot brand "Paibao Robot." The funds will be used to establish a production base in Zhanjiang, Guangdong, and to develop quadruped robots and humanoid robots [14] - "Paibao Robot" aims to enhance its capabilities in commercial robotics and expand its "AI + patrol robot" business model, focusing on overseas market development [14]

Core Viewpoint - Shenzhen Yushu Intelligent Robot Co., Ltd. has undergone significant changes, including an increase in registered capital and the introduction of a new strategic shareholder, Zhiyuan Innovation [1][6]. Group 1: Company Overview - Yushu Intelligent was established in March 2025 as a subsidiary of Yuhua Tian Group, focusing on artificial intelligence applications and intelligent robot development [4]. - The company has increased its registered capital from 50 million to 52.6316 million yuan [1]. Group 2: Leadership and Management - Zhou Ping, Chairman of Yuhua Tian Group, has taken over as Chairman of Yushu Intelligent, bringing extensive management experience [3]. - Yang Bo, Vice President of Yuhua Tian Group, serves as the President of Yushu Intelligent, leveraging his expertise in the smart cleaning robot industry [3]. Group 3: Strategic Direction and Product Development - Yushu Intelligent aims to integrate technology and services, focusing on urban service sectors with a strategy of "operating platform + intelligent equipment + smart services" [3]. - The company is developing various intelligent robots, including unmanned cleaning vehicles and sorting robots, to create comprehensive smart sanitation solutions [3]. Group 4: Strategic Partnerships and Innovations - The partnership with Zhiyuan Innovation aims to establish a technological hub for embodied robotics, enhancing key technology research and data asset development [6]. - The collaboration will lead to the creation of the first "Smart Sanitation Large Model," integrating operational data and intelligent decision-making algorithms for urban service digitalization [6][7]. - Yushu Intelligent has also established the Shenzhen Yushu Embodied Robot Innovation Center, focusing on AI and robotics integration [7].

Core Viewpoint - The article highlights the rapid development and potential of exoskeleton technology in addressing the mobility needs of the elderly population in China, particularly in the context of the 2025 World Robot Conference and the increasing demand for mobility assistance devices [1][3][19]. Group 1: Market Potential and Demand - The 2025 World Robot Conference saw 1.3 million attendees, indicating a strong interest in robotics, particularly exoskeletons, which are expected to see a surge in consumer applications [1][3]. - There are 310 million elderly individuals in China, creating a significant market for mobility solutions like exoskeletons [1]. - The global exoskeleton market is projected to grow from $1.76 billion in 2024 to $30.56 billion by 2032, with a compound annual growth rate (CAGR) of 43.1% [16]. Group 2: Technological Advancements - Traditional exoskeletons faced challenges in adapting to individual users, particularly the elderly, due to fixed parameter control systems [4][5]. - The introduction of AI and deep learning has improved the adaptability of exoskeletons, achieving a 99.2% accuracy in gait prediction [5][7]. - The "踏山AsExo-TK1000" exoskeleton weighs only 2.7 kg, significantly lighter than earlier models, making it more practical for daily use [8][10]. Group 3: Application Scenarios - Exoskeletons are being tested in various scenarios, including tourism and elder care, with successful pilot programs in places like scenic spots and nursing homes [11][13]. - The technology is also being applied in industrial settings, with potential users in high-intensity labor sectors such as manufacturing and construction [13][15]. - In the medical rehabilitation field, exoskeletons can benefit stroke patients and those with partial mobility impairments [15]. Group 4: Accessibility and Pricing - The cost of exoskeletons has decreased significantly, with some products now priced around 9,800 yuan, making them more accessible to the general public [16]. - The company aims to create affordable products to meet the needs of elderly individuals and their families [16][19]. - The establishment of experience centers in multiple cities is part of the strategy to promote awareness and accessibility of exoskeleton technology [19].

2025 世界机器人大会 " 先进机器人与自动化国际学术会议 " 共收到来自中国、法国、美国、新加坡等 10 个国 家的 87 篇论文。经过前期盲审和现场答辩等严格评审流程，评选出 " 最佳学生论文 "—— 《基于深度学习的内 窥镜肺叶切除术手术阶段检测》《基于石墨烯的超灵敏主动脉压力传感器》和 " 最佳会议论文 "—— 《基于相位 表面肌电信号（ sEMG ）的肌肉激活起始点检测》。中国电子学会副理事长兼秘书长、总部党委书记、世界机器 人大会先进机器人与自动化学术会议主席陈英为 " 最佳学生论文 " 和 " 最佳会议论文 " 作者颁奖。 8 月 12 日，为期 5 天的 2025 世界机器人大会在京落下帷幕。从行业应用的精准赋能，到生活场景的智慧融 入；从人形机器人的灵动跃迁，到具身智能的革新突破。这场以 " 让机器人更智慧，让具身体更智能 " 为主题的 国际盛会，通过高水平的论坛研讨、前沿的展览展示、激烈的赛事竞技以及丰富的配套活动，全方位、多维度呈 现了全球机器人领域的最新技术突破、创新应用场景与未来发展趋势，为全球机器人产业高质量发展注入澎湃动 能。 全国政协常委、致公党中央副主席、全国妇联副主席、工 ...

8月8日，2025世界机器人大会（WRC 2025）在北京开幕，"人形机器人第一股"优必选携五大人形机器人亮 相，覆盖工业、商用、科研、陪伴多场景，火力全开，吸引了中共中央政治局委员、北京市委书记尹力，中共 中央政治局委员、北京市委书记尹力，中国科学技术协会主席万钢，工业和信息化部党组书记、部长李乐成， 北京市委副书记、市长殷勇，全国政协常委、致公党中央副主席、全国妇联副主席、中国电子学会理事长徐晓 兰等领导和嘉宾组团参观。 ┃ 五大人形机器人重磅登场，W alker S2和Cruzr S2现场首发 今年 7月，优必选发布全新一代工业人形机器人Walker S2。这款机器人身高1.76米，拥有52个自由度和工业 级灵巧手。通过群脑网络2.0+智能体技术Co-Agent构建AI双循环，Walker S2拥有更智慧的大脑，并具备国内 首个端到端"类人眼"双目立体视觉感知。它采用拟人化设计，可在0-1.8米全空间范围内稳定搬运15公斤负 载，并完成灵活摸地、极限下蹲、远距离抓取等高难度动作，其腰部更可实现±162°灵活旋转。通过刚柔异构材 料的复合应用，Walker S2实现了高强度轻量化的工业级本体设计，能够以2 ...

中关村智友研究院院长王田苗表示，具身智能真正的价值在于能够学习推理然后模仿人，高效地在生产 效率和服务价值上进行匹配。"比如说在服务场景中收拾餐桌，如果人工智能并非机械化的路径进行行 走，而是要先感知，再进行推理，然后形成空间轨迹进而再操作，将大大提高服务价值。" 上海交通大学教授高峰表示，中医有一个词叫"望闻问切"，这四个字其实就是感知，具身智能就是"望 闻问切"的感知产生"神圣工巧"的智能应用。还有一个词叫"看人下菜碟"，这是贬义词，但如果机器人 能"看人下菜碟"，那智商是很高的。 在人工智能与机器人技术快速发展的今天，具身智能正从实验室走向产业化。在8月8日至12日举行的 2025世界机器人大会期间，相关话题引发专家们广泛讨论。 中国工程院院士倪光南表示，机器人智能一般来说可以分为三大部分：脑、眼、行动系统。这三部分组 在一起就变成具身智能系统，能够全面体现机器人的智慧，发挥它的作用。"我们目前的情况是在脑和 眼的方面投入不够，相对来说是一个短板，需要特别予以加强。" 不过具身智能的发展前景是毋庸置疑的，作为机器人的高阶形态和具身智能的重要载体，人形机器人在 生产制造、仓储物流、安全防护、家政服务、教育 ...

【AI产业跟踪】 Qwen 开源 4B 端侧模型 摘要:产业最新趋势跟踪,点评产业最新风向 百度发布智能云数字员工 产业研究中心 | ર્ | 李嘉琪(分析师) | | --- | --- | | క | 010-83939821 | | D | lijiaqi2@gtht.com | | 登记编号 | S0880524040001 | | S | 刘峰(研究助理) | | ર | 0755-23976068 | | E | liufeng6@gtht.com | | 登记编号 | S0880124060013 | 在期回顾 【新材料产业周报】 旭化成与丰田签约向其供应 锂电池隔膜,天羿鹏博等多家新材料企业完成融 O AI 科技前沿 全球首个人形机器人通用视觉感知系统 Humanoid Occupancy 发布 类脑计算机"悟空"发布 全球首个高分辨率三维无线电地图数据集及扩散式建图框架发布 0 风险提示 AI 软件销售不及预期,capex 投资计划变动,AI 产品及大模型研发不及预期 | 往电池间肤,入井鹏将于夕茶制 付在正正成成 资 2025.08.11 | | --- | | 长征十二号运载火箭发射成功 ...

中新社上海8月13日电 (记者 郑莹莹)智元机器人近期对外宣告，其近百台远征A2-W轮式通用机器人已 赴中国一家汽车零部件企业的工厂产线搬箱，实现规模化落地。 伴随具身智能的发展，机器人何时能大批量上岗，深入工厂、迈入家庭？清华大学计算机科学与技术系 教授孙富春在受访时道出人形机器人真正能干活的关键——大的负载能力和拟人的"大脑"。 他认为机器人发展现在进入一个快车道，10年以后，会干活的机器人将在各个行业发挥作用，人形机器 人的形体结构会发生变化，其感知运动能力、认知交互能力、跨场景跨任务能力等都会得到显著提升。 上海飒智智能科技有限公司董事长兼首席执行官张建政告诉中新社记者，机器人在"进化"，要从"婴 儿"时期的慢爬、走路，逐步成长为能力出色的"技术工人"，能精准干活之后，才能适应岗位。以一条 每小时生产90台车的汽车生产线为例，如果机器人发挥不稳定，每次作业有几秒甚至几十毫秒的延迟， 累积起来就会造成生产节奏失序。 今年2月，北京发布《北京具身智能科技创新与产业培育行动计划(2025-2027年)》，计划到2027年，围 绕具身大小脑模型等方面突破不少于100项关键技术，培育千亿级产业集群。 今年3月， ...

Core Viewpoint - The integration of a general brain and general body is the paradigm for constructing general humanoid robots, emphasizing learning from humans as the shortest path to achieve this goal [2][4][15]. Group 1: General Brain and Body Integration - The fusion of a general brain and general body creates a physical world AI evolution flywheel, where a unified model empowers various humanoid robot bodies, adapting to different scenarios and iterating through data feedback [4]. - The ERA-42 general brain model has been released, integrating vision, understanding, prediction, and action, enabling high degrees of freedom in humanoid robot operations through voice commands [4][5]. - The company has developed a fully self-researched supply chain for core components, ensuring high-quality and efficient delivery of products [4]. Group 2: Product Development and Market Position - The company has launched the first full-size bipedal humanoid robot, "Xingdong L7," which surpasses Tesla's Optimus in performance, featuring 55 degrees of freedom and targeting industrial and commercial applications [5][8]. - The "Xingdong Q5" service robot is designed for various scenarios, including retail and healthcare, with over a hundred strategic orders secured from major companies [5][8]. - The company has delivered over 300 products this year, with a significant portion of its market share now coming from overseas, exceeding 50% [8]. Group 3: Software and Hardware Integration - The company has adhered to a "soft and hard integration" approach from the beginning, focusing on developing both the humanoid model and motion control simultaneously [10][11]. - The software is primarily defined by hardware, with models adaptable to various hardware types, ensuring performance optimization [18]. Group 4: Open Source and Industry Trends - The company has released the world's first open-source algorithm for control reinforcement learning, which has gained significant popularity in the humanoid robotics community [20]. - The open-source culture in the AI field has accelerated research and development, leading to a convergence of technology routes [21][22]. Group 5: Future Applications and Market Potential - The company anticipates achieving over 90% efficiency in industrial applications by next year, with ongoing improvements in both software and hardware [26][27]. - The long-term "killer application" for humanoid robots is expected to be in household scenarios, with initial deployments in B2B industrial settings [28][29].