机器人自主性
Search documents
摔倒中摘下VR?特斯拉Optimus的自主性神话破了?
机器人大讲堂· 2025-12-09 10:17
一段 特斯拉的最新 视频 在网上疯传 , 与以往的人形机器人走路炫技不同, 这是一个 "翻车视频"。 在迈阿密炫目的科技演示灯光下,特斯拉人形机器人 Optimus 递出水瓶的瞬间,动作快了零点几秒。水瓶从 它的指间滑落,在地板上弹跳滚动。 紧接着,这个被马斯克称为 " 特斯拉史上最具潜力产品 " 的机器人失去平衡,向后倾倒。 在 即将 摔倒的瞬间 ,它的双臂猛然抬起,双手在空中做出了一个清晰、迅速且无比熟悉的抓握扭转动作。 仿佛在摘下头上的 VR 设备。 现场快门声此起彼伏。这个动作被数十部手机精准捕获,视频当晚就引爆了科技圈。 人们反复观看、逐帧分析: 这到底是 一个人形机器人类人摔倒行为 的笨拙巧合,还是远程操作员在慌乱中的 一次意外 暴露? ▍ Optimus 的 表演失误 ? 马斯克曾激烈反驳关于 Optimus 依赖远程操控的说法。就在 10 月 4 日,他在社交平台 X 上发布 Optimus 学习中国功夫的视频时,还斩钉截铁地回应网友: " AI ,而非遥控操作。 " 然而迈阿密现场那一幕,让这句宣言显得格外微妙。 宣称自主的 机器人 , 那个 " 摘头盔 " 的动作 却 是如此 突兀、 自然 ...
2025WAIC具身智能机器人观察
2025-07-28 01:42
Summary of Key Points from the Conference Call Industry Overview - The conference focused on humanoid robots within the broader artificial intelligence (AI) sector, highlighting a shift in industry development from automation to autonomy [1][2] - The AI field has introduced six new elements: ecology, new scenarios, and capital, which are essential for constructing a commercial closed loop supported by funding [1][4] Core Insights and Arguments - The research goal for humanoid robots has transitioned towards autonomy, emphasizing flexible operations in open physical environments, with capabilities expanding from single-task execution to dexterous movement, long-term reasoning, and collaborative multi-robot tasks [1][6] - The VRA model is still in the exploratory phase for industrial applications, with claims of end-to-end solutions from some companies, but actual implementations rely on Model Predictive Control (MPC) and reinforcement learning [1][7][8] - There is a lack of effective benchmark evaluation standards for general robot models, making it difficult to align objectives and objectively assess technological development levels [1][9] Notable Developments and Innovations - Lenovo introduced a six-legged robotic dog and humanoid robot prototypes for applications in power inspection, forming a commercial ecosystem [3][11] - Haier invested in a new company and showcased its industrial internet platform, promoting the integration of robots into factory-level operations [3][12] - Companies like Cloud Depth and Strong Brain Technology demonstrated advanced robotic capabilities at the conference, including a four-legged robot performing complex maneuvers and a humanoid robot showcasing dexterous hand technology [13] Emerging Trends - The humanoid robot sector is witnessing significant changes, with discussions around foundational models and deployment challenges becoming more prevalent [5][6] - The focus on AI inclusivity and global governance is being emphasized, with China advocating for universal AI applicability to bridge gaps in AI adoption among countries [4] Financial Performance and Market Position - Structural component companies like Henggong Precision are showing strong profitability, having developed key materials for critical components previously reliant on imports [3][16] - Companies such as Rease Smart and Zhaowei Electric are innovating in their respective fields, focusing on lightweight and efficient designs for humanoid robots [17][18] Conclusion - The conference highlighted the dynamic nature of the humanoid robot industry, with significant advancements in technology, funding, and collaborative efforts among companies to address the challenges of autonomy and operational efficiency in real-world applications [1][4][15]