Core Viewpoint - The approval and release of the national standard GB/T 45579-2025 for robot intelligent visual evaluation marks a significant advancement in the global robotics standardization landscape, facilitating the transition of China's robotics industry from "function realization" to "intelligent grading" [1][2]. Group 1: National Standard Significance - National standards play a crucial role in supporting industrial development and promoting technological progress, with a consensus emerging on the need for standard upgrades to drive high-quality economic development [2]. - The new standard addresses the current industry challenge of "having products but no standards," providing a framework for evaluating the intelligent level of robotic products and guiding market development [2]. Group 2: Testing and Evaluation Framework - The standard specifies that robot intelligent visual testing includes visual algorithm testing and complete machine testing, defining 19 key evaluation indicators across five categories: human posture estimation, face recognition, character recognition, object recognition, and safety [2]. - It applies to various types of robots, including industrial, service, and special robots, and innovatively defines three levels of visual intelligence, helping companies accurately position their products and identify development directions [2]. Group 3: Industry Collaboration and Development - Collaborating with leading industry players to establish a unified intelligent evaluation system will enable objective measurement of robotic intelligence levels, enhancing resource utilization and promoting sustainable industry development [3]. - The standardization is essential for driving technological upgrades and compliant market development, with companies like JAKA Robotics leveraging their technological advantages to further the intelligent development of the robotics industry [4]. Group 4: Contributions to Standardization - JAKA Robotics has previously participated in the formulation of multiple national standards, aligning with the "14th Five-Year Plan" for the robotics industry, which emphasizes the role of national technical standards in driving innovation [5][6].

联合创始人兼材料总监朱德航毕业于美国宾夕法尼亚大学，专注于仿生皮肤硅胶材料技术研发，曾荣获中国化 学奥林匹克竞赛一等奖，其在材料科学领域的专业积累为机器人仿生外观的突破提供关键支撑。 机器人大讲堂获悉，超仿生角色人形机器人初创企业「首形科技」近日官宣成功完成 Pre A轮融资。 本轮融 资由招商局创投、深创投共同领投，五源资本、浔商创投跟投，其具体融资金额尚未对外披露。值得注意的 是， 首形科技此前在刚刚成立仅一个月的时候便获得了德迅投资、奇绩创坛、 Taihill、智元机器人的投资支 持。 据机器人大讲堂了解， 首形科技【 全称：首形科技（上海）有限公司 】 成立于 2024年6月， 是一家专注 于高端超仿生人形机器人研发的企业。公司 致力于将尖端人工智能技术与类人机器人技术深度融合至服务机 器人领域，打造兼具高度交互能力与逼真外观的人形机器人产品。 作为类人机器人领域的探索者，首形科技以 "引领世界走向更智能、更和谐的未来"为愿景，持续深耕人工智 能与机器人技术的核心领域，涵盖电机驱动、大型语言模型、机器人运动控制、表示学习、自我监督学习及革 命性仿生皮肤技术等方向。通过技术创新，公司旨在推动机器人智能化 ...

Core Viewpoint - The article emphasizes the significant impact of artificial intelligence (AI), particularly generative AI, on various industries and the increasing demand for AI-related content and research. Springer Nature is actively embracing these trends by publishing insightful research and developing AI tools and services to enhance academic and industrial collaboration [1]. Group 1: AI Resources and Services - Springer Nature is set to launch a new collection of AI books in 2025, covering topics from machine learning algorithms to neural network research impacting healthcare. This collection aims to provide a comprehensive resource for researchers, educators, and professionals [2]. - The AI book collection includes diverse resources such as monographs, conference proceedings, textbooks, and manuals, catering to different levels of readers and facilitating a deeper understanding of the rapidly evolving AI field [2]. Group 2: Core Research Areas - The AI book collection focuses on core research areas, including machine learning, deep learning, natural language processing, computer vision, and AI ethics, among others [3]. Group 3: AI-Driven Services - Springer Nature offers AI-driven services, including text and data mining (TDM) tools that help researchers extract valuable information from large datasets, enhancing their ability to discover patterns that traditional analysis may miss [46]. - The Methods Muse platform simplifies experimental design and optimization for life sciences research, integrating with protocols.io to streamline the research process [50]. - Nature Research Intelligence provides data-driven insights to help academic, governmental, and corporate researchers make informed decisions, track research trends, and enhance their academic impact [53][55].

1、 微型机器人有望用于治疗鼻窦炎 来自香港中文大学与广西、深圳、江苏、扬州和澳门大学学者的最新合作研究发表在《科学机器人》杂志 上。研究人员称，可以部署大量微型机器人，用于治疗顽固的鼻窦感染，机器人可以通过纸巾吸入鼻子， 因为这些微型机器人的宽度仅为人类头发的几分之一，目前该研究已在临床前试验中成功将其插入动物鼻 窦。机器人群落通过穿过鼻孔的导管注入鼻窦腔，并通过电磁力引导至目标，在那里它们会被加热并催化 化学反应，从而消灭细菌感染。人们希望这种精准靶向技术最终能够减少人们对抗生素和其他通用药物的 依赖。专家认为，这些药物有望在五到十年内应用于临床治疗膀胱、肠道和鼻窦感染。 2、 NAU研究人员发布开源外骨骼框架 北亚利桑那大学（NAU）的研究人员希望未来残疾人能够借助机器人腿自行行走。该大学发布了一款开源 机器人外骨骼，以加速其研发。这个名为OpenExo的开源系统提供了构建单关节或多关节外骨骼的全面说 明，包括设计文件、代码和分步指南。任何人都可以免费使用。研究人员表示，目前，开发外骨骼机器人 复杂的机电系统成本高昂且耗时，这很可能导致许多研究在启动之前就夭折。但北亚利桑那大学副教授扎 克·勒纳的生物机 ...

水下激光雷达领域的技术探索已历经多年积淀，但长期以来，市场主流产品 仍 受困于 "体积庞大"与"成本 高昂 "的双重瓶颈，这 导致其应用范畴 长期停留在科研 和 工业级 层面， 在日常消费场景及商业领域的规 模化应用始终难以推进。 近日， 思岚科技正式官宣，推出业内首款消费级水下激光雷达 ——RPLIDAR 家族全新成 员 RPLIDAR U1 ，成功打破水下激光雷达应用领域壁垒， 将水下激光雷达带入消费级应用 新赛道 。 RPLIDAR U1 尺寸只有乒乓球大小，可以非常方便的安装在各类设备当中。 同时， 通过技术创新，其成本 大幅优化， 可满足消费类产品的使用。 凭借小型化设计和成本优化优势 ， RPLIDAR U1 不仅填补了消费级 水下激光雷达的市场空白，更推动 了 水下场景迈入高精度激光 SLAM 导航的全新时代 ， 或将重构水下智 能设备的技术应用生态 。 | 项目 | 参数 | 备注 | | --- | --- | --- | | 水下探测 | 最大:5米 | 最大值在浊度 <0.5 NTU 且 不存在强 环 境 | | 距离 | 典型:2米 | 光的条件测得 | | 典型精度 | <4cm | ...

因为 A 股市场为了保护股民利益收紧，这两年来港交所这次捡了不少漏，尤其对现金流较差早期融资节奏太 快的企业，因为资金链和盈利性问题，只能选择跑到港股上市，这使得 A 股错失不少机器人企业。 为了抢救市场， 6 月 18 日，在 2025 陆家嘴论坛上证监会主席 吴清 表示，创业板正式启用第三套标准， 支持优质未盈利创新企业上市。将继续充分发挥科创板示范效应，加力推出进一步深化改革的 1+6 政策措 施，重启未盈利企业适用科创板第五套标准 ( 允许符合国家战略、突破关键核心技术且市场认可度高的企 业，在无盈利情况下以预计市值 40 亿元门槛上市 ) 上市 。 证监会表示在持续抓好 " 科创板八条 " 落地实施的基础上，以设置专门层次为抓手，重启未盈利企业适用科 创板第五套标准上市，推出一揽子更具包容性、适应性的制度改革，着力打通支持优质科技型企业发展的堵点 难点，同时进一步加强投资者合法权益保护。一方面，设置科创板科创成长层，在科创成长层的定位、企业入 层和调出条件、强化信息披露和风险揭示、增加投资者适当性管理等方面明确具体要求。另一方面，围绕增强 优质科技型企业的制度包容性适应性，推出 6 项改革举措。 值 ...

Group 1 - Foxconn and NVIDIA Corporation announced a collaboration to deploy humanoid robots in Foxconn factories for the production of NVIDIA's AI servers, marking the first time NVIDIA products will be produced with robotic assistance [1] - The deployment of humanoid robots is expected to be completed in the coming months, with the robots set to begin operations in the first quarter of next year [1] Group 2 - UR Robotics launched UR Studio, an online simulation tool built on the advanced PolyScope X software platform, allowing users to interact with UR's robot products and components [4] - UR Studio is designed to ensure solutions adapt to real environments, facilitating smoother deployment and faster return on investment, while also identifying potential issues early to reduce downtime [4] Group 3 - Google DeepMind is releasing an offline version of its Gemini Robotics AI model, which is capable of performing various physical tasks without specialized training [7] - The model can adapt to new situations with only 50 to 100 demonstrations and has been trained on Google's ALOHA robot, with potential applications on different types of robots [7] Group 4 - Optimotive introduced the Iris MK2, a rugged autonomous robot designed for challenging terrains, capable of performing automated 3D scanning and remote inspections [10] - The Iris MK2 can support terrain measurement before construction and preventive maintenance checks during construction phases [10] Group 5 - Apptronik established Elevate Robotics, a wholly-owned subsidiary focused on automating industrial tasks beyond human capabilities [13] - Elevate Robotics will operate independently and concentrate on developing and commercializing multi-purpose automation solutions related to heavy industrial-scale operations [13]

Core Viewpoint - The article highlights the significant milestone achieved by Wuhan Huawike Intelligent Technology Co., Ltd. (HIT Huawike) with the mass production of electronic skin for humanoid robots, marking a new era in the development of sensitive robotic components [2][22]. Group 1: Electronic Skin Technology - The electronic skin developed by Huawike enables robots to have a sense of touch, allowing them to perform delicate tasks such as gently squeezing soft candy without damaging it [5][6]. - The technology integrates nearly 100 tactile points within a square centimeter, providing robots with the ability to detect weight changes as small as 1 gram, equivalent to the weight of a business card or a feather [8][9]. - The electronic skin technology has evolved from laboratory experiments to mass production, with a capacity of 2 million automotive-grade units, indicating a robust supply chain for humanoid robots [13][14][22]. Group 2: Production and Innovation - Huawike's production line operates at a speed of 30 meters per minute, achieving a precision of 0.01 millimeters, which is crucial for the accurate arrangement of tactile sensors [16][19]. - The company employs a "Roll-to-Roll" printing process, which significantly reduces manufacturing costs and time, facilitating the large-scale application of electronic skin [18][19]. - Huawike has developed a complete sensory system for humanoid robots, extending the electronic skin technology from hands to the entire body, enhancing robots' ability to perceive and interact with their environment [12][24]. Group 3: Industry Impact and Collaboration - The mass production of electronic skin is expected to eliminate a critical supply barrier for the humanoid robot industry, enabling more domestic robots to be equipped with advanced sensory capabilities [22][24]. - Huawike has established deep collaborations with local and national leading humanoid robot manufacturers, ensuring rapid technology iteration and standard-setting in tactile perception [24][26]. - The company has participated in over 8 industry standard-setting initiatives and holds more than 100 intellectual property rights, positioning itself as a leader in the flexible intelligent sensing field [21][22]. Group 4: Global Expansion - Huawike is actively expanding its presence overseas, showcasing its products at international exhibitions such as Automatica in Munich, indicating a strong competitive stance in the global market [28][27]. - The company aims to redefine global standards for precision gripping through its innovative electronic skin technology, contributing to the advancement of flexible sensing technology [28].

自然界中的水生生物因其生活环境和捕食方式的差异，演化出了多种多样的游动策略。在这些策略中， 头足 类动物，例如鱿鱼和鹦鹉螺，通过从其腔体中快速喷射水流来实现快速游动，从而具有游动速度快、能量效率 高等优势， 同时具备静音、结构简单、适应环境广等特点。 但人工模仿这种高效推进方式存在挑战：目前尚缺乏一种能像生物肌肉一样在水下同时实现高驱动应变、高驱 动力和高驱动速度的人工肌肉，因而无法有效驱动封闭性柔性腔室快速喷射水体，完成高效喷水推进。 ▍突破水下驱动瓶颈，研发新型仿生喷射机器人 近期的研究发现，将液晶弹性体 (LCE)纤维通过绳结结构编织成肌肉后，其在水下甚至深海环境中的驱动性 能可获得显著提升，并表现出优异的力学响应和环境适应性。然而，这种人工肌肉在水下的 驱动速率 仍显不 足，单靠其自身的牵拉作用难以实现对腔室的快速挤压，因此尚难满足高频、高流量喷射推进的严苛要求。 针对这一问题，来自北京大学的刘珂研究员团队进行了深入研究，并提出了一种实现水下仿生喷射的可缩放机 器人设计方案。 该机器人的推进力源于一种创新设计策略，集成了导电绳结人工肌肉、仿折纸软壳和机载控 制模块。基于液晶弹性体的导电绳结人工肌肉在水 ...

Core Viewpoint - Huawei is advancing its embodied intelligence strategy through the launch of the CloudRobo platform, focusing on software and hardware ecosystem development while partnering with various companies to enhance industrial applications and market reach [1][26][27]. Group 1: CloudRobo Platform and Technology - The CloudRobo platform is based on the Pangu large model, featuring multi-modal capabilities and a "cloud-network-end" collaborative architecture, which includes the "robotic three-piece set" [2]. - The "embodied brain" (planning model) can perform complex task planning with over 10 steps, ensuring smooth production processes in manufacturing environments [2]. - The "embodied small brain" (execution model) specializes in high-precision control, achieving over 90% success rates in complex tasks [2]. - The "digital universe workshop" generates 80% of training data through simulated environments, requiring only 20% real data for validation [2]. Group 2: Strategic Partnerships - Huawei Cloud emphasizes a "be integrated" strategy, focusing on providing cloud-based intelligence while leaving the physical robot development to partners, aiming to integrate intelligence into various industries [3]. - Efort Robotics showcased a voice-interactive sorting workstation, developed in collaboration with Huawei Cloud, enhancing task execution efficiency [5][6]. - Daotong Technology presented a smart inspection solution that integrates drones and ground robots, achieving nearly 90% accuracy in oilfield inspections [8]. - Yijiahe Group introduced a humanoid robot for healthcare, leveraging Huawei Cloud's AI capabilities for enhanced service delivery [10][11]. - XCMG highlighted its autonomous platform robot, which can operate in various intelligent modes for complex environments [13]. Group 3: Industry Applications and Innovations - TuoStar is developing humanoid robots for logistics, integrating Huawei's AI technology to improve operational efficiency [15]. - Leju Robotics launched the first 5G-A embodied intelligent robot, demonstrating capabilities in high-concurrency scenarios [17]. - Qianxun Intelligent showcased its humanoid robot Moz1, which utilizes a multi-modal model for real-time task execution [19]. - Youai Zhihui presented solutions for semiconductor manufacturing, achieving significant efficiency improvements in logistics and operations [22]. - Xinghai Map demonstrated a teleoperation platform that enhances data collection and training for embodied intelligence applications [24]. Conclusion - The launch of the CloudRobo platform marks a significant step in Huawei's embodied intelligence strategy, focusing on collaboration with partners to drive the robotics industry towards larger-scale implementation [26].