Workflow
机器人大讲堂
icon
Search documents
单笔10亿融资频现!9月机器人行业融资环比激增40%,自变量机器人、首形科技引领具身智能融资热潮
机器人大讲堂· 2025-10-08 09:49
根据机器人大讲堂 &立德智库不完全统计,2025年9月,中国机器人行业融资活动呈现显著回暖态势,共发 生融资事件56起,较8月的39起环比增长超40%,市场活跃度明显提升。融资总额预估达到54亿-78亿元, 较8月28亿-47亿元的区间实现大幅跃升,反映出资本对机器人赛道,尤其是头部项目和前沿技术领域的出手 力度与信心显著增强。 从融资规模来看,大额融资事件表现尤为抢眼。 本月亿元级融资高达 13起,不仅在数量上近乎8月(7起) 的两倍,其战略意义更为深远。其中, 通用具身智能模型研发领域出现单笔高达 10亿元的A+轮融资 ,显示 出市场对这一前沿制高点的强烈追捧; 服务机器人 领域亦有龙头企业完成约 10亿元融资 ,凸显了其在商用 清洁等成熟场景中实现规模化落地的巨大潜力与商业价值。这些大额融资清晰地表明,资本正以前所未有的力 度,向那些具备定义下一代机器人平台潜力的关键技术节点集中。 融资轮次分布延续了早期为主的格局,但中后期融资活力增强 。 种子 /天使轮与A轮系列融资合计占 比达到 86 %,是创新活力的主要来源。与此同时,B轮及以后的中后期融资事件数量有所增加,如昂泰微精等企业 获得持续资本加持,反 ...
刚登Nature 子刊!华科大柔性 FEbots 机器人,突破微型机器人集成瓶颈
机器人大讲堂· 2025-10-07 04:05
Core Insights - The article discusses the development of micro-robots with insect-level autonomy and flexibility, aimed at performing tasks in complex environments such as disaster recovery and industrial inspections [1] - A research team from Huazhong University of Science and Technology has created a flexible electronic robot (FEbot) that integrates structural adaptability and embedded computing capabilities [1][3] FEBots Design and Modular Architecture - The FEbot utilizes a modular architecture combined with an oscillatory driving mechanism, differing from traditional robots that rely on precise leg trajectory control [3] - The FEbot consists of two main components: a programmable flexible electronic module and a distributed setae array made from superelastic alloys [5] Oscillatory Drive and Asymmetric Friction - The movement of the FEbot is driven by an oscillatory mechanism that relies on asymmetric friction, allowing for net forward displacement during each cycle [7] - The study identified three key states in the movement cycle, with the optimal configuration achieving a peak speed of 109.5 mm/s at a driving frequency of 450 Hz [9] Multi-Modal Movement and Environmental Sensing - FEbots exhibit remarkable multi-modal movement capabilities, allowing for quick reconfiguration based on task requirements [10] - The I-type design, inspired by centipedes, can navigate tight spaces and carry loads up to 5.1 times its weight, while the II-type design showcases omnidirectional movement [11][12] Embedded Artificial Intelligence - The research team incorporated an embodied intelligence framework into the FEbot, enabling real-time intelligent decision-making through hyperdimensional computing [15] - The FEbot demonstrated autonomous behaviors such as danger avoidance and thermal gradient tracking, showcasing its ability to process multiple sensor inputs simultaneously [18]
00后黄彦钧再出圈!央视报道“手搓”真·钢铁侠!网友:军迷减速带!
机器人大讲堂· 2025-10-07 04:05
大家还记得去年 靠 自制兽装控制器 火遍 全网 的重庆交大学生黄彦钧吗? 没错,就是那个实现了 生物仿生学眨眼 、配备 夜视仪 和 图像识别算法 ,还在头套里装了 换气系统 的 " 硬核玩 家 ", 冲上了热搜。 这回,他又整活了。 最新作品是 一套 真· " 钢铁侠 " 战甲! 央视亲自转发,外网直接炸锅了。 这套战甲 全身能自动开合,迅速变形贴合身体, 例 如 肩周全自由度, 手臂 可自动闭合。 有趣的是, 机甲还接入了AI语音对话, 和贾维斯助理对话起来真的超酷好么! 还配有 机甲 信息终端,看信息就和看手表一样简单。 最让网友惊讶的是,IMU体感触发系统可以根据特定动作实现 手 臂 电磁冷发射架的启动, 能发射"导弹",根据展示需要,加装烟雾、激光束等武器道具! 更厉害的是, 战甲 中的 助力外骨骼还可以提供额外动力,增强人体极限! 网友直呼:" 现实版钢铁侠照进现实 。 " " 这不 军迷 减速带吗! " 这位 00 后此前还研发过 外骨骼辅助行走机器人 并获得专利,参与过 清华大学脑机接口领域实验室 项目 ,甚至以 " 战术装备研发人员 " 的头衔登上央视军事 频道。 从 " 拆家电少年 " 到 ...
绳驱外骨骼Kinethread亮相,动觉反馈步入轻量化、低成本时代
机器人大讲堂· 2025-10-06 04:05
Core Viewpoint - The article discusses the development of Kinethread, a full-body haptic exoskeleton system designed to enhance immersive experiences in virtual and augmented reality by providing realistic tactile feedback, overcoming the limitations of current consumer-grade and professional devices [1]. System Design: Rope-Driven Mechanism and Modular Structure - Kinethread utilizes a computer-controlled electric reel and rope transmission mechanism, combined with a modular tactical vest for load distribution, enhancing comfort during prolonged wear [2]. - The system integrates 10 electric reels, with 8 positioned around the waist to optimize ergonomic design and effectively hide reverse tension during operation, allowing clearer perception of force feedback [2]. Motor Selection and Weight Considerations - The system employs two types of high-torque brushless DC motors, with the larger motor weighing 265 grams and providing a maximum pull of 42N, while the smaller motor weighs 140 grams with a maximum pull of 17N [4]. - The total weight of the system, including the battery, is kept below 5 kilograms, with a wear time of under 30 seconds, balancing performance and practicality [4]. Pulley Mechanism and Force Output - The design incorporates traditional mechanical structures like pulleys to change force direction and provide mechanical advantage, achieving a maximum output force of 120N through a three-pulley structure on the back [6]. - The system's design allows for dual-direction drive from a single motor, reducing complexity and overall weight [6]. Control and Software Capabilities - Each motor connects to a custom driver, with multiple microcontrollers communicating with the host via USB, utilizing the open-source SimpleFOC library for closed-loop control at approximately 2kHz [11]. - The software, developed on Unity, supports various haptic feedback modes, allowing for simultaneous simulation of vibrations and dynamic forces, enhancing the immersive experience [11]. Performance Metrics - Kinethread demonstrates potential for consumer applications with performance tests showing peak force achieved within 50 milliseconds and a vibration feedback bandwidth of 200Hz, capable of producing over 3N of force at 100Hz [12]. - The system's standby power consumption is 1.5W, with typical scene power consumption averaging 10.9W, allowing for over 6 hours of wireless operation with a 12V 5600mAh lithium battery [12]. Application Scenarios - Kinethread can simulate various tactile effects, including static weight, dynamic forces, impact forces, and social interactions, enhancing user experience in virtual environments [13]. User Experience Evaluation - A user study with 12 participants indicated that the combination of vibration and dynamic feedback significantly outperformed other conditions in terms of realism, immersion, and instinctive reactions, highlighting the importance of dynamic feedback [15]. Future Prospects - Future developments aim to address limitations such as the inability to simulate pushing forces and potential risks associated with cable routing, while exploring additional types of force feedback and optimizing motor configurations for improved realism [16].
清华、北信科、复旦团队解读具身智能!大语言模型与世界模型如何让机器人懂物理、会思考?
机器人大讲堂· 2025-10-06 04:05
当机器人能像人类一样理解自然语言指令,还能预判环境变化、自主规避物理风险时,通用人工智能的落地似 乎不再遥远。 近日,清华大学计算机科学与技术系,北京信息科学与技术国家研究中心,复旦大学可信具身 智能研究所联合发布《 Embodied AI: From LLMs to World Models》。 系统 性 梳理了具身智能的技术 脉络,尤其聚焦大语言模型与世界模型的协同 。 ▍ 先搞懂什么是具身智能?它和普通 AI 有啥不一样? 单模态与多模态具身智能 纯视觉的机器人,在昏暗环境或动态场景里很容易迷路;纯语言控制的机器人,可能会因为没考虑物理规律提 出离谱指令。 | | | Embodied AI: From LLMs to World Models | | | --- | --- | --- | --- | | EAI § II | EAI with LLMs/MLLMs § III | EAI with WMs § IV | EAI with MLLMs and WMs § V | | The Historical View § II-A | LLMs Boost EAI § III-A | WMs Bo ...
全球眼科手术机器人技术趋势全景:高精度突破、AI 具身智能演进与行业整合方向
机器人大讲堂· 2025-10-05 04:06
众所周知,眼科手术长期以来对精度和稳定性有极高要求。视网膜膜剥离、玻璃体切除、视网膜注射等操作对 外科医生的手部控制能力提出了微米级的挑战,传统手术中难以避免的手颤和视觉–动作不一致增加了手术风 险。 眼科手术机器人的出现,为上述难题提供了全新的技术解决方案。通过高刚性的机械臂结构和精密的运动控制 算法,机器人系统能够有效滤除人手固有震颤,实现亚毫米级甚至微米级的定位稳定性。此外,手术机器人的 应用也显著提升了操作的精细度与可重复性。这不仅降低了因操作误差导致的并发症风险,也使一些以往被视 为过于复杂或风险极高的眼底手术成为可能。 近年来,随着人工智能技术,特别是以视觉-语言-动作协同为代表的具身智能(Embodied AI)的快速发 展,眼科手术机器人正迎来新一轮技术变革。 未来的机器人将不再仅仅是外科医生的"放大镜"和"稳定器", 而是逐步具备环境感知、决策分析与自主执行能力的手术伙伴。例如,在图像识别和路径规划能力支持下,机 器人有潜力实现从入眼定位到目标区域操作的全过程自主化这种高度自主化不仅将进一步提升手术效率与一致 性,更可能为眼科手术的标准化、规模化普及开辟全新路径,尤其在医疗资源相对匮乏的地区 ...
撒贝宁把机器人梗玩出续集!
机器人大讲堂· 2025-10-05 04:06
这个国庆假期, 中央广播电视总台大型 季播节目 《城市风华录》深圳篇 播出, 节目由 总台主持人 撒贝宁、王嘉宁、孟湛东与同济大学特聘教授诸大建、演员 王鸥组成 的 " 风华团 " ,带着对 " 奇迹 " 的追问走进深圳。 在 9 月份的节目录制期间,就有深圳网友在龙岗偶遇撒贝宁,当时他正在机器人 6S 店 体验各种 " 黑科技 " 。节目播出后,网友们纷纷表示: " 小撒这次是彻 底和机器人杠上了。 " 节目中的一大亮点是撒贝宁来到位于深圳龙岗的 全球首家机器人 6S 店, 一进门就问出一句让现场气氛瞬间活跃的话: " 你是真人还是机器人? " 现场嘉宾被逗 乐,网友直呼 " 综艺感满分 " 。 从元宵晚会模仿机器人的经典场面,到如今深圳探店, 这 "六亲不认的机器人步伐梗 " 也是被撒老师玩出续集了! 回到前面提到的外骨骼, 佩戴者穿上这套外骨骼机器人 可轻松提起 20 公斤的哑铃 。 据介绍,这款设备已经从最初的 " 爬山神器 " 进化成多场景应用的实用工 具,在 养老助行、工业防护等领域 都有广泛应用。 在店内,撒贝宁体验了被称为 " 赛博外挂 " 的外骨骼机器人, 穿上后迈出了六亲不认的步伐 。 咖 ...
千寻智能高阳团队最新成果:纯视觉VLA方案从有限数据中学到强大的空间泛化能力
机器人大讲堂· 2025-10-04 04:05
Core Viewpoint - The article discusses the introduction of a new strategy called "State-free Policy" in visuomotor control for robots, which enhances spatial generalization capabilities by removing state information from the input, relying solely on visual observations [1][10][24]. Group 1: State-free Policy Overview - The State-free Policy allows robots to perform tasks effectively even when the training data is strictly controlled, such as fixed desktop height and object positions [3][10]. - This policy is based on two key conditions: representing actions in a relative end-effector space and ensuring comprehensive visual input for task observation [11][13]. Group 2: Experimental Results - Extensive experiments demonstrated that State-free Policy significantly improves spatial generalization, achieving a success rate of 0.98 in height generalization and 0.58 in horizontal generalization for the pen insertion task [14][24]. - In challenging tasks like folding clothes and fetching bottles, State-free Policy outperformed state-based models, showcasing superior horizontal generalization capabilities [17][20]. Group 3: Advantages of State-free Policy - State-free Policy exhibits higher data utilization efficiency, maintaining performance even with limited data, unlike state-based strategies that tend to overfit [20][21]. - The policy also shows advantages in cross-platform adaptation, requiring less adjustment compared to state-based strategies, leading to faster convergence and higher success rates [21][22]. Group 4: Sensor Design Considerations - The research suggests reevaluating sensor designs, particularly the necessity of overhead cameras, as they may introduce performance issues due to changes in object positions [22][23]. - The findings indicate that using dual wide-angle wrist cameras can provide sufficient task observation without the overhead camera, maintaining high success rates in various scenarios [23].
瞩目!一日狂揽顶刊5篇+1封面!中国团队硬核构筑柔性电子强国之路
机器人大讲堂· 2025-10-04 04:05
近日,国际知 名顶刊《Science Advances》上演了一场令人瞩目的"中国时刻"—— 同一天内,5篇来自中 国科研团队的柔性电子领域重磅研究集中登刊,其中1篇更直接拿下当期封面! 放眼全球科研圈,如此高密 度、高含金量的成果爆发,实属罕见。 作为一门新兴交叉科学与技术,柔性电子凭 "轻薄柔透 "的核心特性,在 航空航天、公共安全、国防军工、 健康医疗等国计民生关键领域 展现出广阔的应用前景,并为培育新质生产力提供了重要技术支撑。特别是在 未来信息芯片、高端智能装备、新型电子器件等战略方向 ,该技术有望催生突破性创新,开拓全新的产业空 间。预计到2028年,中国制造的柔性电子在泛物联网领域的应用规模将突破 3000万美元 ,预计占未来10 —15年柔性电子整体市场的 40%份额 ,将逐步发展为国家战略性新兴产业的重要支柱,在未来产业格局中 占据主导地位。 下面为大家具体呈现这 5 篇同一日登上《Science Advances》顶刊的中国团队重磅研究。 ▍清华大学团队在磁驱动柔性电池集成机器人上取得新进展 近日,清华大学深圳国际研究生院智能感知与机器人(Smart Sensing and Robotic ...
从演奏《青花瓷》到《我和我的祖国》!国内这四家机器人乐队都这么先进了?
机器人大讲堂· 2025-10-04 04:05
2025 年世界机器人大会期间,亦庄机器人产业园的机器人研究院聚集了大量观众。除了餐厅机器人服务员在 现场服务,机器人乐队的现场演奏更是让其成为餐厅里最亮眼的部分,键盘手机器人指尖在琴键上飞速跳跃, 吉他手机器人的拨弦动作流畅自然,整个乐队配合默契,奏响的乐曲既饱含科技感又充满感染力。那些曾经仅 存于科幻场景下的机器人乐队,如今正逐步走进现实。 ▍ 机械臂弹扬琴、灵巧手敲架子鼓!第三代和璇机器人乐队成 2025WRC显眼包 今年 8月 在北京举行的 2025 世界机器人大会上,来自杭州的第三代和璇机器人乐队成为全场焦点。在机器 人城市会客厅的体验空间内,三位机器人乐手分别执掌钢琴、扬琴与架子鼓,默契合奏《青花瓷》等曲目,悠 扬旋律与精准演奏吸引众多观众驻足鼓掌,更获央视新闻频道《新闻 1+1》等主流媒体重点关注。 这 个备受瞩目的机器人乐队 , 来自 杭州海创人形机器人创新中心。该创新中心由浙江大学与杭州市余杭区 联合共建,与浙江大学机器人研究院实行 "两块牌子、一套班子" 运营模式,自 2024 年 8 月揭牌以来,便 以场景智能为核心推进技术研发与产业化落地。此次亮相的第三代和璇机器人,正是其科艺融合探索 ...