仿生学作为连接自然与工程的前沿交叉领域，正不断推动智能机器人系统的革新。鱼类在复杂水下环境中展现 出的卓越机动性和感知能力，启发了仿生机器鱼的设计与发展。然而，在水下环境光照不足、声学干扰频繁等 恶劣条件下，传统视觉与声呐传感技术常常面临性能瓶颈。相比之下，鱼类独有的 侧线系统 ——能感知周围 水流速度和压强变化——为水下机器人提供了全新的感知范式。 近年来，研究人员开发了人工侧线系统（ Artificial Lateral Line System, ALLS），用于模拟鱼类的感知 能力，广泛应用于流场感知、避障和多鱼协同等任务。但若要实现真正的自主智能，机器人还必须 准确估计 自身的运动状态 ，如速度、摆动模式和轨迹变化。对于自由游动的仿生机器鱼来说，因其推进依赖周期性摆 动，所引发的强流固耦合问题使得传统建模手段难以适用，而现有的深度学习方法虽具精度，往往缺乏可解释 性与泛化能力。 ▍ 提出新框架，突破仿生机器鱼感知难题 针对这一挑战，来自 北京大学先进制造与机器人学院智能仿生设计实验室的谢广明教授团队 ， 提出了一种 融合模态分解与物理建模的可解释、可泛化的数据驱动自运动状态估计框架。 该 方 法 基 ...

当前， 神经系统疾病 已成为 全球第二大死亡原因 。近年来， 脑肿瘤、帕金森病、强迫症、癫痫等常见脑 神经 功能性重大疾病随年龄增长患病率显著上升，严重影响患者生活质量。目前临床常用的诊疗手段包括穿 刺活检、深部电极植入、射频消融、肿瘤切除和局部给药等。 传统框架式立体定向手术依赖术前影像与框架匹配实现手动 靶点定位 ，操作复杂且对医生经验要求较高；而 无框式非磁共振兼容立体定向机器人手术虽通过机械臂与术前影像实现 了自动靶点定位给 ，却因缺乏术中实 时影像与颅内环境监测能力，难以应对术中脑组织漂移等问题。 术中磁共振影像 （ MRI） 引导技术可在手 术过程中同步实现解剖结构与功能评估，动态适应组织形变并避开关键功能区，但其临床应用仍面临多自由度 大范围运动与高精度定位难以兼顾、受限空间下长距离精准穿刺操作复杂，以及术中磁共振成像时空分辨率难 以同时优化等挑战。 此外，强磁场环境对机器人系统的材料、驱动与传感技术提出了严格的磁兼容性要求。 在这一背景下， 国家重点研发计划 "智能机器人"专项设立"脑神经介入核磁兼容穿刺机器人技术与系统"项 目。 该项目面向脑疾病精准治疗的迫切临床需求，聚焦驱动、感知、成像导 ...

近日美国具身智能公司Physical Intelligence旗下的VLA模型π0.5正式开源，π0.5最核心的能力在于，通过异 构数据协同训练与多模态数据融合，基于优化的模型架构，实现了机器人在复杂现实场景下强大的泛化能力， 使其能理解任务语义、拆解复杂任务流程并精准执行动作。 ▍ π0.5开源模型都有哪些技术亮点？ π0.5 的一大技术亮点是采用了异构数据协同训练的方式。该模型整合了来自多个机器人、高级语义预测、网 络数据等多种不同来源的数据，通过协同训练，能够让模型能够实现更广泛的泛化，更好地适应现实世界中的 机器人操作任务。在训练过程中，模型不仅能学习到物理技能的执行方法，还能理解每个技能背后的语义背 景，推断任务的高级结构，甚至可以从其他机器人那里借鉴物理行为经验。 此外π0.5能够将图像观察、语言命令、目标检测、语义子任务预测和低级动作等多模态数据示例进行融合。 这些数据不是简单地叠加在一起，而是在训练中深度融合。比如，图像观察数据让机器人能识别环境中的物体 信息，语言命令数据帮助机器人理解人类意图，目标检测数据能让机器人快速锁定任务相关物体，语义子任务 预测数据可辅助规划任务流程，低级动作数据则 ...

Core Viewpoint - Zhengyu Industrial has announced the establishment of Shenzhen Feibosen Robot Technology Co., Ltd., marking its entry into the "electronic skin" sector, focusing on flexible tactile sensing technology for robots [1][2]. Company Overview - Zhengyu Industrial, founded in 1998, is one of the early companies in China to produce automotive suspension system shock absorbers, with main products including automotive suspension system shock absorbers, rubber damping products, and engine sealing components [4]. Financial Performance - In the mid-2025 financial report, Zhengyu Industrial reported a revenue of 1.356 billion RMB for the first half of the year, representing a year-on-year growth of 39.62%. The net profit attributable to shareholders was 119 million RMB, a significant increase of 420.67% year-on-year [6]. New Market Entry - The establishment of Feibosen Company aims to accelerate Zhengyu Industrial's strategic layout in embodied intelligence and industrial sectors, ultimately maximizing investment value [2]. Technical Leadership - The company has brought in Dr. Dong Bo, a dual PhD holder, as a core technical leader for Feibosen. Dr. Dong has extensive experience in optical devices and has published over 140 academic papers and filed more than 30 patents [8][11]. Market Potential - The flexible tactile sensing system, commonly referred to as "electronic skin," can simulate human skin perception and has applications in medical monitoring, smart wearables, robotics, and virtual reality. The market for electronic skin is gaining attention as embodied intelligence scenarios mature [2]. Competitive Landscape - The global flexible pressure sensor market is projected to reach 6.8 billion USD by 2033, with a compound annual growth rate (CAGR) of 12.1% from 2025 to 2033. Several listed companies, including Fulei New Materials, Jinghua New Materials, Hanwei Technology, GoerTek, and Shenhao Technology, are competing in the electronic skin market [12][18].

机器人大讲堂获悉，「如身机器人」近日官宣完成数千万元人民币天使 +轮融资 ， 本轮由道彤投资独家投资 完成，其所融资金将主要用于核心技术迭代、产品工程化落地、养老场景试点推进以及市场化前期部署。 据机器人大讲堂了解， 如身机器人（全称：上海如身机器人科技有限公司） 成立于 2023年6月，是一家专 注于智慧康养场景的具身智能机器人企业，致力于开发可实际进入家庭、服务于老年群体的智能康养机器人。 该公司聚焦康复健身与家庭服务两大领域，以模块化、高性价比及人机协作为技术方向，推动康复机器人产品 的研发与应用。通过自主研发的模块化力控关节和趣味化游戏内容，如身机器人构建了一套覆盖全康复周期与 多类场景的康复机器人产品矩阵，并计划逐步拓展至健身支持与家庭日常服务领域。目前，如身机器人总部设 于上海张江，在德国慕尼黑设有欧洲研发与运营中心，并在浙江嘉兴建有生产基地。 在核心团队方面 ， 如身机器人 创始成员主要来自德国顶尖人工智能机器人院士研究所、港股上市公司微创 机器人以及欧洲独角兽企业 Agile Robots等行业领先机构。 公司 首席科学家为中国工程院外籍院士。创始人兼 CEO师云雷博士持有天津大学机械工程硕士 ...

Core Viewpoint - The article highlights the upcoming 25th China International Industry Fair, featuring an "Industrial Inspection" activity that allows participants to visit manufacturing sites and witness the implementation of smart manufacturing solutions [1][22]. Group 1: Industrial Inspection Activity - The event will take place from September 23 to September 26, 2025, with over 1000 professional manufacturing-related participants expected [1][22]. - Free shuttle buses will be provided to transport participants from the exhibition to the factories, facilitating efficient resource connections [1]. Group 2: Featured Companies and Locations - New Times is a leading company in China's industrial robot market, with a payload range of 1-600KG and expertise in key technologies such as control systems and software [5]. - The company was recognized as a national intelligent manufacturing demonstration unit in 2021 and is a member of the Haier Kaos industrial internet ecosystem [5]. - Advantech's Kunshan Manufacturing Park features smart monitoring and operations across its four production areas, having received multiple awards for its green and intelligent manufacturing practices [9]. - Inovance Technology's Industrial Control Museum showcases the evolution of automation technology and the company's contributions to industrial civilization [12]. - Panasonic's Suzhou facility focuses on the development and manufacturing of general sensors and automation solutions, with a significant portion of its sales coming from the Chinese market [16]. Group 3: Registration and Participation - Registration for the inspection activity is free but requires prior approval to ensure participants are industry professionals [19][20]. - The deadline for registration is September 19, 2025, and notifications will be sent within three working days after the deadline [21].

更炸裂的是，大会联合四大机器人展团 ——张江机器人谷、漕河泾新兴技术开发区、合肥科大硅谷、成都人 形机器人创新中心，直接打造了一个 "机器人小镇" 。 40家知名具身智能企业集体亮相，包括青龙、智元、开普勒、傅利叶、卓益得、青心意创等本体公司，帕西 尼、无限工坊、爱观视觉等上下游企业 ， 展示人机共生的科技发展图景。 在 "具身智能：从泛化到行动，重塑产业未来"论坛上，大家最关心的问题是： 机器人啥时候能真正走进千家 万户？ 继 7月世界人工智能大会（WAIC）之后，上海再次迎来机器人大集结！ 9月10日—13日，" 2025Inclusion·外滩大会 "在上海黄浦世博园区举行。这个全球科技与金融领域的重要 对话平台，今年搞了个大的： 10000平米科技展览、5000平米科技集市，近200家企业参展，超30项科技 新品亮相。 ▍ 机器人 "新物种"集体亮相， 大会成 具身智能大秀场 走进外滩大会的 "机器人小镇"，仿佛进入了一个未来世界的生活场景。 开普勒 K2大黄 蜂 这个 行业领先落地的 "蓝领人形机器人"已经实现"充电1小时，连干8小时"的超长续航， 还能双臂协同搬运30公斤重物。现场观众排队跟它握手 ...

近日，特斯拉动作频频，先是发布"宏图计划"第四篇章完整文件，将人工智能与人形机器人置于战略核心，然 后又是疑似Optimus新型号人形机器人曝光，金色涂装+人机对话+精巧拟人的灵巧手引发热议，随后又开通 了"TeslaAI"微博账号，马斯克又在最近的All-In上表示特斯拉目前正在设计的Optimus V3将解决手部灵活 性、拥有AI大脑，并且有望实现量产。 特斯拉的这一系列消息，无疑再度引发了市场对其人形机器人发展的关注。特斯拉在人形机器人领域的持续投 入——Optimus的迭代与量产，离不开其供应链体系。 本文盘点了部分有望在特斯拉供应链上发力的企业。 ▍ 三花智控 浙江三花智能控制股份有限公司成立于1994年公司秉持"专注领先、创新超越"的经营战略，以热泵技术和热 管理系统产品的研究与应用为核心，专注于冷热转换、温度智能控制的环境热管理解决方案开发，致力于建筑 暖通、电器设备和新能源汽车热管理领域的专业化经营。 三花智控自成立以来，始终专注于热管理技术的研发、推广和应用，业务主要分为制冷空调电器零部件业务和 汽车零部件两大板块。 制冷空调电器零部件业务方面，三花智控主要从事控制元器件、子系统和技术解决方 ...

海洋覆盖了地球表面 70%以上的面积，蕴藏着丰富的资源。为更好地开发利用海洋，操作简便、高效且自主 性强的水下机器人成为关键装备。机器鱼在一众水下机器人中，因其出色的灵活性与环境友好性，近年来受到 广泛研究关注，并被应用于多个水下领域。 然而，水下环境的复杂性对机器鱼的感知能力提出了严峻挑战，尤其是流速感知能力，直接影响到其环境适应 性与任务执行效果。受自然界鱼类侧线器官的启发，此前已有研究人员开发了压力型、变形型等多种人工侧线 （ ALL）传感器。但现有ALL传感器在实际应用中仍面临显著干扰：一方面易受机器鱼自身运动（如偏航、 俯仰）影响，另一方面也受到鱼体周围涡流的干扰。 因此，如何开发一种能够实现特定方向流速感知、并有效抑制多方向噪声的 ALL传感器，已成为当前机器鱼 研究领域的重要课题。 ▍提出新型传感器，助力实现精准流速感知 面对上述挑战， 浙江大学赵朋教授与张承谦博士课题组开展了深入研究，并提出一种基于磁触觉感知原理、 具备流速解耦检测能力的磁式 ALL传感器。 前不久，该研究成果的 相关论文已以 " Online velocity estimation of a robotic fish usi ...

Core Viewpoint - In August 2025, China's robotics industry financing activities showed a phase adjustment after a high in July, with a total of 39 financing events, a 22% decrease from July's 50 events, indicating a need for improved market transparency [1][2]. Financing Overview - The estimated financing amount in August ranged between 2.8 billion to 4.7 billion yuan, significantly down from July's 8.4 billion to 9.8 billion yuan, reflecting a tightening of capital outflow [1]. - There were 7 financing events exceeding 100 million yuan, with notable amounts such as 500 million yuan for Mech-Mind Robotics and 300 million yuan for Fourier Intelligent Technology, indicating continued focus on leading companies with clear technological barriers and commercialization capabilities [1][2]. Financing Round Distribution - Early-stage financing saw a significant increase, with seed/angel rounds reaching 19 events and A rounds totaling 15 events, together accounting for 87% of the total, up from 66.7% in July [2]. - Mid-to-late stage financing showed structural adjustments, with only 6 events in B rounds and beyond, indicating a high financing threshold and a market trend of "early-stage activity and mid-to-late stage differentiation" [2]. Sector Performance - The humanoid robots and embodied intelligence sector stood out with 14 financing events, estimated to raise 960 million to 1.54 billion yuan, making up about one-third of the total industry financing in August [4]. - Capital is actively investing in both "hardware-software synergy" and "AI large model empowerment," focusing on leading manufacturers like Fourier and Songyan Power, as well as companies like Ruoyu Technology that provide foundational AI large models [4]. Key Companies - **Fourier Intelligent Technology**: Completed a 300 million yuan E+ round investment, marking its entry into the unicorn club with a pre-investment valuation of 8 billion yuan. The funds will support the development of humanoid robot production lines [8][11]. - **Songyan Power**: Announced a multi-hundred million yuan A++ round financing, led by Jinpu Investment, focusing on breakthroughs in motion control and environmental interaction technologies [15][17]. - **Mech-Mind Robotics**: Secured 500 million yuan in D round financing to enhance its embodied intelligence technology and expand its product line and application scenarios [19][23]. - **SpeedBot Robotics**: Completed a 100 million yuan B+ round financing to further develop industrial embodied intelligence products and expand into maritime and lithium battery industries [24][26]. Conclusion - Overall, the financing market in August 2025 for China's robotics industry exhibited characteristics of "total decline, early-stage strengthening, and structural optimization." Capital is increasingly concentrating on projects with core barriers and clear commercialization prospects, indicating a rational layout and value exploration strategy in the long term [7].