当前人形机器人正逐步进入实际应用部署阶段，2025年被行业普遍视为其量产化元年。随着市场需求的快速 增长，人形机器人本体对传感、边缘计算、电源及连接系统的需求也将同步攀升。作为全球领先的半导体制造 商，ADI正将自动化与机器人领域视为未来的 "第二增长曲线"。据其规划，到2030年ADI的自动化业务收入有 望翻一番；尤其在人形机器人赛道，单台设备搭载的 ADI 芯片价值可能达到数千美元，这一数值是当前AMR 的十倍。 ▍ ADI联手NVIDIA押注人形机器人赛道 根据2024年行研数据，全球模拟芯片市场规模约为850-900亿美元，预计未来5年将以 5%-7%的年均增速增 长，其中汽车电子、工业控制是主要增长动力。整个市场存在高度集中的现状，前五大厂商占据近60%的份 额。其中2025年ADI在全球模拟芯片市场的份额为13%。主要用于工业、汽车信号链芯片领域，市场份额仅 次于德州仪器（TI）。在数据转换器市场，ADI市场份额高达40%，远超竞争对手两倍，占公司营收30%以 上。 从产业来看，全球前五大模拟芯片厂商德州仪器、ADI、英飞凌、意法半导体、美信更多集中在工业控制、汽 车电子、消费电子、医疗设备等领域 ...

在现代机器人技术的浪潮中，我们似乎总离那个理想中的 "家庭机器人"还差一步。它们能看、能听，甚至能 跑能跳，却依然难以在真实的家庭、商场或街道中像人类一样应对自如。哪怕硬件再先进、感知再敏锐，一旦 置身于开放、多变的环境中，机器人往往显得笨拙、迟疑，甚至 像是 " 社恐 "附体。 而最近，字节跳动 Seed 团队推出的 Robix ，试图从根本上重新设计这个"大脑"。 究其原因，并不全在 机器人的 "手脚"，更在于 其 "大脑"。 当前大多数机器人系统采用分层架构：高层负责思考（比如用大语言模型做任务拆解），底层负责执行。问题 就在于，这些所谓的 "大脑"——比如大语言模型（ LLM ）或视觉 - 语言模型（ VLM ） ， 往往只擅长分 解任务，却在面对物理世界的空间关系、位置感知，或是人类的实时打断、多轮对话时，表现得捉襟见肘。 ▍ Robix 统一模型架构：告别 " 机械式应答 " 更直白地说，现有的机器人系统缺乏一种真正意义上的 "智能"：它们无法像人一样，在动态环境中同时进行 理解、推理、回应和行动。 Robix 作为 一个统一的视觉 - 语言模型架构 ， 与传统分层系统将机器人推理、任务规划和自然 ...

值得一提的是，今年3月该公司曾完成 2300 万美元种子轮融资，当时由硅谷 VC CRV 和 First Round Capital 领投、真格基金参投，估值约1亿美元。时隔仅半年，公司估值已上涨6倍。 Dyna Robotics于2024年9月在美国得克萨斯州休斯敦成立，在上海设有硬件研发中心，国内主体公司上海达 纳灵动科技有限公司于2025年2月完成注册。并同期开放多个硬件研发岗位招聘。 ▍ 三位联创履历靓眼 其中一位浙大校友 中美两地协同研发 Dyna Robotics拥有三位创始人。其中，联合创始人兼 CEO Lindon Gao 与联合创始人兼 CTO 杨世远 （York Yang）是连续创业者，二人于 2016 年共同创办智能购物车公司 Caper AI，该公司在 2021 年以 3.5 亿美元的价格被 Instacart 收购。杨世远还凭借突出的技术能力与创业成果，入选 2020 年北美福布斯 30U30 榜单。值得一提的是，杨世远是浙江大学2010 级信息与通信工程专业本科，辅修竺可桢学院 ITP （创新与创业管理强化班），后获得 UCLA 计算机专业硕士。 近日，硅谷初创具身智能企业Dyna ...

Core Viewpoint - The article emphasizes the importance of embodied intelligence as a key link between AI capabilities and real-world value, as outlined in Huawei's "Intelligent World 2035" report, which envisions a future where over 90% of Chinese households will have smart robots within the next decade [1][19]. Group 1: Strategic Foundation of Embodied Intelligence - Huawei defines embodied intelligence as a physical AI carrier that integrates various technologies, including visual, tactile, language, and action models (VTLA), perception interaction, computing storage, communication networks, and energy technologies [3][6]. - Embodied intelligence serves as a crucial bridge connecting virtual cognition and physical action, essential for the development of Artificial General Intelligence (AGI) [4][3]. Group 2: Methodology for Implementation - Huawei's approach to embodied intelligence involves a layered technical architecture and domain-specific evolution, focusing on foundational technology, scenario validation, and ecosystem collaboration [7][10]. - The foundational technology architecture consists of six core modules, including perception, decision-making, action, and support, with specific breakthrough directions and metrics for each module [8]. Group 3: Competitive Strategy - Huawei aims to avoid a technology parameter race in the global competition for embodied intelligence, instead focusing on ecosystem collaboration, data differentiation, and end-cloud integration to build unique competitive advantages [11][13]. - The company recognizes the need for vertical data platforms tailored to industry-specific needs, enhancing the precision of embodied intelligence in vertical fields [13]. Group 4: Boundaries and Limitations - Huawei acknowledges the technical and commercial limitations in the development of embodied intelligence, particularly in fine manipulation and tactile perception, which remain significant challenges [14][16]. - Ethical and safety risks associated with physical interactions between embodied intelligence and humans are highlighted, necessitating the establishment of unified interaction standards and responsibility frameworks [16][14]. Group 5: Long-term Vision - Huawei's strategy for embodied intelligence reflects a long-term commitment to integrating technology into daily life, emphasizing the balance between technological breakthroughs and ecological maturity [19]. - The company envisions a future where embodied intelligence serves humanity, enhancing quality of life and operational efficiency across various sectors [19].

参赛者需要完成一系列模拟日常生活的任务，如处理食物、开关门、使用工具等，这些都是截肢者日常生活中 常遇到的挑战。 2024 年 10 月，来自 HANDSON 团队的徐敏在苏黎世赛场上创造了历史。她 不仅完成了所有任务，还是 唯一成功完成最难题 "触觉袋"任务的选手——需要仅凭假手从盲取袋中取出具有特定形状和柔顺性的物体。 近日 ，国际顶级期刊 《 Science Robotics 》 刊登了一篇令人瞩目的研究，详细记录了 东南大学宋爱国教 授领导的 HANDSON 团队在 2024 年 Cybathlon 全球辅助技术大赛中夺冠的技术细节。这标志着中国在辅 助技术领域的研究首次获得如此高级别的国际学术认可。 该成果以 " Arm prosthesis with dexterous control and sensory feedback delivers winning performance at Cybathlon "发表在《 Science Robotics 》上。 科学与技术的完美结合让 47 岁的徐敏——一位失去右上肢三十年的女性 ， 在这场被称为 "仿生学奥运 会"的赛事中，以惊人表现击败了来自 ...

Group 1 - Humanoid, a London-based robotics and AI company, launched the HMND 01 Alpha prototype on September 18, which stands 87 inches tall, can travel at 4.4 mph, and carry a payload of 33 lbs [2] - The HMND 01 Alpha is the UK's first humanoid robot designed for industrial use, aimed at performing complex tasks in logistics, retail, and manufacturing [2] - The robot was developed in just seven months and is powered by NVIDIA's Jetson Thor platform, enabling autonomous thinking and action [2] Group 2 - OpenMind announced the release of its open-source operating system, OM1, which allows anyone to program robots, described as the "Android for robots" [5] - OM1 provides developers with necessary components and frameworks to create, test, and deploy robotic applications without being tied to a specific platform [5] - The system integrates various AI models for advanced reasoning capabilities, enabling rapid prototyping of voice-controlled robots and drones [5] Group 3 - Siemens Healthineers and Stryker are collaborating to develop a neurosurgical robot that can perform selective and emergency neurovascular interventions [8] - The robot aims to improve surgical precision and reduce treatment time, which is critical for patient outcomes in conditions like acute ischemic stroke [8] Group 4 - Norwegian startup Sonair raised $6 million to expand the distribution of its ADAR (Acoustic Detection and Ranging) sensors, designed for robots operating around humans [11] - The ADAR system uses 3D ultrasound technology for real-time spatial awareness and has been certified for safety standards [11] - Sonair plans to use the funding to increase shipments to manufacturers in Asia, Europe, and North America, addressing the growing demand for autonomous robots in logistics and manufacturing [11] Group 5 - Icarus, a startup based in New York, secured $6.1 million in seed funding to develop a workforce of space robots equipped with mechanical arms and claws [14] - The company's first robot will handle unpacking and loading cargo, focusing on logistics tasks in space rather than scientific experiments [14] - Icarus plans to conduct flight tests next year and demonstrate the robot's capabilities on the International Space Station [14]

Core Viewpoint - The article discusses the emergence of "soft human-machine interfaces" based on flexible electronic technology, which aims to enhance the interaction between humans and machines through intuitive and natural means. The technology faces challenges such as accurately interpreting physiological signals and achieving cost-effective, scalable manufacturing [1][2]. Group 1: Flexible Electronic Technology - A research team from Shanghai University of Science and Technology has developed a printed human-machine interface that includes electronic skin for surface electromyography (sEMG) collection and feedback, multimodal tactile sensing soft robots, and machine learning algorithms for gesture classification and material recognition [2]. - The core breakthrough of this technology is the electronic skin (e-skin), a thin electronic sensing device that can be attached to the human body to monitor various physiological signals in real-time [5]. Group 2: Production Techniques - The research team utilized an efficient integrated printing technology, including direct ink writing (DIW), infrared laser engraving, and laser cutting, to achieve large-scale production of multi-material, high-density sensor arrays [7]. - Various functional inks, such as silver ink, carbon ink, and PDMS/C, were used to print electrical circuits as narrow as 40 micrometers on flexible substrates [7]. Group 3: Intelligent Algorithms - The challenge of enabling machines to accurately understand human intentions is addressed through an adaptive machine learning method that combines linear mapping networks (LMN) and initial time models (ITM) [11]. - The LMN adjusts the weights of signals from different channels to adapt signals from various users to a unified standard distribution, while the ITM captures local features in time series with low latency and high accuracy [12]. Group 4: Multimodal Sensing - The soft human-machine interface integrates a "sensory system" for robots, allowing them to recognize object characteristics through touch by incorporating temperature, pressure, thermal conductivity, and electrical conductivity sensors [14]. - The pressure sensor features a capacitive design with a sensitivity of 10.5 pF/kPa, maintaining stable performance over 2000 tests, while the combination of thermal and electrical conductivity sensors improved material recognition accuracy from 63.99% to 98.03% [16]. Group 5: Application Prospects - This technology has broad application prospects, establishing a complete interactive ecosystem that includes signal collection, intention recognition, action execution, and sensory feedback, forming a closed-loop human-machine interaction cycle [18]. - In the medical field, it offers new hope for upper limb amputees, achieving an average accuracy of 94.36% in recognizing 11 hand and finger gestures, even with significant time delays and reduced signal strength [18].

微小型移动机器人作为新兴研究领域，因其结构紧凑、机动性强等特点受到广泛关注。目前已有电磁电机、压 电致动器、磁致动器等多种驱动方式的微型机器人被开发。其中，压电机器人因响应速度快、精度高、结构紧 凑等优势，成为工业与生物医学应用中具有潜力的技术方向。 ▍提出3DMPPR，实现耦合干扰最小化 面对上述挑战，来自 北京大学的研究人员 基于结构动力学建模和结构设计，设计出一款三自由度小型压电平 板机器人（ 3DMPPR）。 然而，该领域仍面临多项技术挑战。其中基于行波驱动的压电平板机器人虽具备运动平稳、结构简单、易于承 载和易于批量生产等优点，但在实现多自由度运动方面存在明显局限。现有设计大多难以在单一平板结构中同 时生成正交或旋转行波，导致运动自由度受限，直线运动性能不佳、转弯半径过大等问题普遍存在。此外，振 动模态耦合现象干扰运动精度与方向控制，而通过独立多驱动单元拼装、高阻尼材料连接或附加支撑结构等传 统解决方法，又易导致结构复杂、装配困难及制造成本上升。 研究团队首先建立了行波驱动下平板机器人在动态接触摩擦中的动力学模型 ， 揭示了激励参数对振幅、切向 速度和动摩擦力的影响机制，从而确定了最大化驱动力的最优 ...

今日，国产灵巧手赛道头部企业灵巧智能重磅出击，向外界分享其在机器人灵巧操作领域的最新研究成 果， 正式发布灵巧智能DexCanvas数据集 。该数据集规模达20TB，包含1000小时真人操作记录，是涵盖多模态 人手操作数据的重磅资源，将为机器人灵巧操作领域注入强劲动力。 ▍成本、规模、真实性难以兼得，具身智能数据采集困境待解 当前 AI在物理世界的应用中，虽已实现理解人类语言、识别物体与场景、规划任务步骤等能力，但 在物理世 界中的 "最后一公里"，即让机器人像人类一样灵活地抓握、理解语言、识别物体和场景、感知并调节力度、 适应不同物体等方面，仍是一个待突破的难题。 这一瓶颈很大程度上源于当前大规模、高质量、多模态交互 数据集的缺乏。 一般来说，机器人在实际场景中的操作表现往往受到感知不确定性、动力学复杂性和环境变化敏感性的制约， 也因此 数据集的规模与质量直接决定了模型在真实环境中的表现。 从技术实现路径来看，具身智能操作的数 据采集方式目前主要为遥操作、 视频学习和仿真合成。 遥操作通过专业设备记录人类专家的动作和力控信息，能获得高质量、高精度的真实数据，尤其适合精密力控 任务，但存在设备昂贵、效率低以及 ...

在人形机器人领域，谐波关节作为关键部件，其性能与设计直接影响着机器人的灵活性、负载能力和成本效 益。随着人形机器人应用场景的不断拓展，从工业生产到服务领域，市场对谐波关节的需求也在持续增长。 然而，当前市场仍面临着诸多挑战。一方面，传统谐波关节的重量较大，限制了机器人的负载能力和灵活性， 尤其是在轻量化设计需求日益凸显的当下，这一问题尤为突出。另一方面，现有解决方案在力位控制精度、安 装便捷性以及成本控制等方面存在明显不足，用户在实际应用中常常面临精度不够、安装复杂和成本过高的困 境。这些问题不仅制约了人形机器人在更多领域的广泛应用，也反映出市场对更高效、更灵活、更经济的谐波 关节解决方案的迫切需求。 基于对行业趋势的敏锐洞察和对市场痛点的深度剖析，国产谐波减速器及机电一体化执行器领域头部企业同川 精密积极投身研发，并于近日对外官宣 将在第二十七届中国国际工业博览会（ CIIF 2025）期间正式推出 HarmoCore 系 列 轻 量 化 谐 波 关 节 。 据 机 器 人 大 讲 堂 了 解 ， HarmoCore 系 列 涵 盖 TCHL-11M 、 TCHL- 14M、TCHL-17H、TCHL-20 ...