2026-07-17 证据索引
每条声明均绑定可定位的精确原文和内容哈希。
创造人机竞速新历史——人形机器人马拉松见证中国未来产业奔跑之路01a628ded1826714
声明：从区域布局看，产业集聚效应正在持续显现。在北京亦庄，300余家人形机器人企业，基本构建起覆盖核心零部件、整机到应用场景的全产业链体系，为技术创新与场景应用深度融合提供可靠支撑；在深圳南山，人形机器人从研发到产业化的全链条布局已经形成。企业与科研机构紧邻联动，上下楼就是上下游，将近一半的零部件可在方圆10公里内配齐供应，极大提升了产业协同效率与产业化速度。
从区域布局看，产业集聚效应正在持续显现。在北京亦庄，300余家人形机器人企业，基本构建起覆盖核心零部件、整机到应用场景的全产业链体系，为技术创新与场景应用深度融合提供可靠支撑；在深圳南山，人形机器人从研发到产业化的全链条布局已经形成。企业与科研机构紧邻联动，上下楼就是上下游，将近一半的零部件可在方圆10公里内配齐供应，极大提升了产业协同效率与产业化速度。
Home | Robotics: Science and Systems04c6301c32afee8e
声明：The website of the RSS 2025 conference can be found here.
The website of the RSS 2025 conference can be found here.
千寻智能完成5.28亿元Pre-A轮融资，领先VLA模型加速具身智能落地0867dd0d12a90720
声明：千寻智能（Spirit AI）成立于2024年2月，是国内领先的AI+机器人全栈生产力级技术能力的具身智能公司。公司团队成员来自UC Berkeley、CMU、清华大学、北京大学等顶尖高校，以及字节、小米、腾讯等知名企业。成立短短一年多，千寻智能凭借卓越的技术实力和创新理念，已成功完成多轮融资，并致力于打造世界级通用人形机器人和下一代具身大模型与学习算法，让通用的机器人伙伴走进千家万户，驱动世界迈向智能机器人时代。
千寻智能（Spirit AI）成立于2024年2月，是国内领先的AI+机器人全栈生产力级技术能力的具身智能公司。公司团队成员来自UC Berkeley、CMU、清华大学、北京大学等顶尖高校，以及字节、小米、腾讯等知名企业。成立短短一年多，千寻智能凭借卓越的技术实力和创新理念，已成功完成多轮融资，并致力于打造世界级通用人形机器人和下一代具身大模型与学习算法，让通用的机器人伙伴走进千家万户，驱动世界迈向智能机器人时代。
Qwen - Alibaba Cloud09011dadd561d216
声明：Qwen - Alibaba Cloud
Qwen - Alibaba Cloud
Inside X Square Robot’s Bold Plan for Real Home Robots095e97ad765d3984
声明：It reports that this reaches performance comparable to an all-robot dataset at roughly a 20-fold lower cost of collection, driven mainly by how much cheaper the wearable rig is than a teleoperation setup.
It reports that this reaches performance comparable to an all-robot dataset at roughly a 20-fold lower cost of collection, driven mainly by how much cheaper the wearable rig is than a teleoperation setup.
Publications09791636fd92b75c
声明：* Equal contribution/co-first author.
* Equal contribution/co-first author.
Scaling Behavior Foundation Model for Humanoid Robots0a08ab5a5a432d23
声明：In this work, we revisit the scaling recipe for BFMs and demonstrate that substantial performance gains can be achieved through the coordination of three core components: 1) the learning paradigm of motion tracking that reformulates diverse humanoid control problems as the reproduction of integrated whole-body behaviors in the global frame; 2) the strategic synergy between on-policy rollout quantity and reference motion diversity; and 3) the expressive and scalable model architecture termed Humanoid Transformer that facilitates the natural emergence of structured behavioral representations.
In this work, we revisit the scaling recipe for BFMs and demonstrate that substantial performance gains can be achieved through the coordination of three core components: 1) the learning paradigm of motion tracking that reformulates diverse humanoid control problems as the reprod…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
isaac-sim/IsaacLab v2.3.20b2cfbcd9382494a
声明：Related PR: #3287 **New Features:** - **Permanent Wrench Composer**: Applies forces/torques that persist across simulation steps until explicitly changed - **Instantaneous Wrench Composer**: Applies forces/torques for a single simulation step, then automatically resets - **Composability**: Multiple forces and torques can now be added together on the same body - **Mixed Frame Support**: Seamlessly compose local and global frame wrenches **Migration Guide:** **Old API (Deprecated):** ```python # Old method - overwrites previous forces asset.set_external_force_and_torque( forces=torch.ones(1, 1, 3), torques=torch.ones(1, 1, 3), body_ids=[0], env_ids=[0], is_global=False, ) ``` **New API:** ```p
Related PR: #3287 **New Features:** - **Permanent Wrench Composer**: Applies forces/torques that persist across simulation steps until explicitly changed - **Instantaneous Wrench Composer**: Applies forces/torques for a single simulation step, then automatically resets - **Compos…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
近百名玩家涌入具身数据：一年融资44.7亿，谁能真靠“卖数据”赚钱？0b2df0040932bbb6
声明：原因有两方面：外部，真机数据和人类数据的供给快速增加，价格持续下降，仿真数据的规模和成本优势被削薄；内部，sim2real gap仍没有太好的解法，很难高保真还原真实世界中的摩擦、形变、力觉与触觉反馈。
原因有两方面：外部，真机数据和人类数据的供给快速增加，价格持续下降，仿真数据的规模和成本优势被削薄；内部，sim2real gap仍没有太好的解法，很难高保真还原真实世界中的摩擦、形变、力觉与触觉反馈。
A Steerable Model with Emergent Capabilities0b3d6544532aded4
声明：A Steerable Model with Emergent Capabilities
A Steerable Model with Emergent Capabilities
Qwen - Alibaba Cloud0c98e29b78d7e825
声明：The latest Qwen3 models adopt hybrid thinking modes ("Thinking" and "Non-Thinking") that allow you to flexibly control reasoning performance, speed, and costs.
The latest Qwen3 models adopt hybrid thinking modes ("Thinking" and "Non-Thinking") that allow you to flexibly control reasoning performance, speed, and costs.
NVIDIA/Isaac-GR00T n1.6.1-release0d43ab1c08aac00d
声明：Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_6/ Minor Fixes for GR00T 1.6 Release - Removal of dangling references in Dockerfile and pyproject.toml's.
Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_6/ Minor Fixes for GR00T 1.6 Release - Removal of dangling references in Dockerfile and pyproject.toml's.
isaac-sim/IsaacLab v3.0.0-beta2.patch10fcdcd74817abc98
声明：# Isaac Lab 3.0 Beta 2 - Patch 1 This is a small patch release on top of the previous Isaac Lab 3.0.0 Beta 2 release, including an update to support Isaac Sim 6.0.1, which includes fixes and improvements for NuRec workflows (https://docs.isaacsim.omniverse.nvidia.com/6.0.1/overview/release_notes.html).
# Isaac Lab 3.0 Beta 2 - Patch 1 This is a small patch release on top of the previous Isaac Lab 3.0.0 Beta 2 release, including an update to support Isaac Sim 6.0.1, which includes fixes and improvements for NuRec workflows (https://docs.isaacsim.omniverse.nvidia.com/6.0.1/overvi…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
Hugging Face model lerobot/lingbot_va_robotwin1052c4c16f29acaf
声明：Hugging Face model lerobot/lingbot_va_robotwin
Hugging Face model lerobot/lingbot_va_robotwin
NVIDIA/Isaac-GR00T n1.6.1-release181bb3465cbab8c6
声明：NVIDIA/Isaac-GR00T n1.6.1-release
NVIDIA/Isaac-GR00T n1.6.1-release
NVIDIA/Isaac-GR00T n1.6.1-release19def6909db62a7a
声明：Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_6/ Minor Fixes for GR00T 1.6 Release - Removal of dangling references in Dockerfile and pyproject.toml's.
Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_6/ Minor Fixes for GR00T 1.6 Release - Removal of dangling references in Dockerfile and pyproject.toml's.
isaac-sim/IsaacLab v2.3.219ec039584c77fdf
声明：isaac-sim/IsaacLab v2.3.2
isaac-sim/IsaacLab v2.3.2
huggingface/lerobot Release v0.5.11ac1b2793dc7d002
声明：huggingface/lerobot Release v0.5.1
huggingface/lerobot Release v0.5.1
《人形机器人创新发展指导意见》解读1b03fc4e023e72cc
声明：《人形机器人创新发展指导意见》解读
《人形机器人创新发展指导意见》解读
CosFly-VLA: A Spatially Aware Vision-Language-Action Model for UAV Tracking1b6e8f887012b6c9
声明：Relative to OpenVLA, CosFly-VLA-0.8B reduces open-loop Average Displacement Error (ADE) by 34.1% on seen-test and 35.3% on unseen-test.
Relative to OpenVLA, CosFly-VLA-0.8B reduces open-loop Average Displacement Error (ADE) by 34.1% on seen-test and 35.3% on unseen-test.
Qwen - Alibaba Cloud1d5669cc3b5b3932
声明：The latest Qwen3 models adopt hybrid thinking modes ("Thinking" and "Non-Thinking") that allow you to flexibly control reasoning performance, speed, and costs.
The latest Qwen3 models adopt hybrid thinking modes ("Thinking" and "Non-Thinking") that allow you to flexibly control reasoning performance, speed, and costs.
Hugging Face dataset agibot-world/AgiBotWorldChallenge-20261defd67d722b9778
声明：Tags: license:cc-by-nc-sa-4.0, region:us
Tags: license:cc-by-nc-sa-4.0, region:us
《人形机器人创新发展指导意见》解读1e2fe5231d29658e
声明：《指导意见》从特种领域、制造业典型场景、民生及重点行业三类方向提出意见措施。一是加快人形机器人在特种环境应用，面向恶劣条件、危险场景作业等需求，强化复杂环境下本体控制、快速移动、精确感知等能力。二是聚焦3C、汽车等制造业重点领域，提升人形机器人工具操作与任务执行能力，打造人形机器人示范产线和工厂，在典型制造场景实现深度应用。三是拓展人形机器人在医疗、家政等民生领域服务应用，满足生命健康、陪伴护理等高品质生活需求，推动人形机器人在农业、物流等重点行业应用落地，提升人机交互、灵巧抓取、分拣搬运、智能配送等作业能力。
《指导意见》从特种领域、制造业典型场景、民生及重点行业三类方向提出意见措施。一是加快人形机器人在特种环境应用，面向恶劣条件、危险场景作业等需求，强化复杂环境下本体控制、快速移动、精确感知等能力。二是聚焦3C、汽车等制造业重点领域，提升人形机器人工具操作与任务执行能力，打造人形机器人示范产线和工厂，在典型制造场景实现深度应用。三是拓展人形机器人在医疗、家政等民生领域服务应用，满足生命健康、陪伴护理等高品质生活需求，推动人形机器人在农业、物流等重点行业应用落地，提升人机交互、灵巧抓取、分拣搬运、智能配送等作业能力。
千寻智能完成5.28亿元Pre-A轮融资，领先VLA模型加速具身智能落地1e84e832f3d3aa91
声明：千寻智能完成5.28亿元Pre-A轮融资，领先VLA模型加速具身智能落地
千寻智能完成5.28亿元Pre-A轮融资，领先VLA模型加速具身智能落地
Hugging Face dataset agibot-world/AgiBotWorldChallenge-20262179f3ed331534c3
声明：Tags: license:cc-by-nc-sa-4.0, region:us
Tags: license:cc-by-nc-sa-4.0, region:us
近百名玩家涌入具身数据：一年融资44.7亿，谁能真靠“卖数据”赚钱？22bdbb3db718a4e9
声明：比如，曾以仿真数据为核心的光轮智能，也开始采集人类数据；曾是最坚定的仿真派之一的银河通用，今年6月发布了全身遥操作系统，拥有了遥操数据采集能力。
比如，曾以仿真数据为核心的光轮智能，也开始采集人类数据；曾是最坚定的仿真派之一的银河通用，今年6月发布了全身遥操作系统，拥有了遥操数据采集能力。
RoboTTT: Context Scaling for Robot Policies2365ce5c1fb1eb1e
声明：RoboTTT: Context Scaling for Robot Policies
RoboTTT: Context Scaling for Robot Policies
Moravec's Paradox and the Robot Olympics246f2b9bb25eea7b
声明：We did not do everything possible for the highest success rate (as discussed, e.g., in our recent work on using RL for optimizing reliability and speed), and the policies for these tasks are often not consistent, though on average they have a success rate of 52% and a task progress of 72%.
We did not do everything possible for the highest success rate (as discussed, e.g., in our recent work on using RL for optimizing reliability and speed), and the policies for these tasks are often not consistent, though on average they have a success rate of 52% and a task progre…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
Qwen - Alibaba Cloud253df30662fd1e17
声明：Multimodal models, including Qwen-VL (large language vision model), Qwen-TTS (text-to-speech model), and Qwen-Audio, support cross-modal processing, while the latest Qwen3-Omni is capable of processing and reasoning across text, audio, and vision modalities simultaneously.
Multimodal models, including Qwen-VL (large language vision model), Qwen-TTS (text-to-speech model), and Qwen-Audio, support cross-modal processing, while the latest Qwen3-Omni is capable of processing and reasoning across text, audio, and vision modalities simultaneously.
Hugging Face dataset agibot-world/AgiBotWorld202626304e729acaa537
声明：Tags: task_categories:robotics, language:en, license:cc-by-nc-sa-4.0, size_categories:1K<n<10K, modality:image, modality:text, region:us, agibot, imitation-learning, embodied-ai, lerobot, real-world, dual-arm
Tags: task_categories:robotics, language:en, license:cc-by-nc-sa-4.0, size_categories:1K<n<10K, modality:image, modality:text, region:us, agibot, imitation-learning, embodied-ai, lerobot, real-world, dual-arm
How Cheap Ground Robots Are Rewriting Frontline Warfare in Ukraine27c882a1993e8bd2
声明：The Droid TW 12.7, by DevDroid, is shown here outfitted with a 0.50-caliber M2 Browning machine gun that can be aimed and fired by a remote operator using a tablet and an encrypted communications link.DevDroid
The Droid TW 12.7, by DevDroid, is shown here outfitted with a 0.50-caliber M2 Browning machine gun that can be aimed and fired by a remote operator using a tablet and an encrypted communications link.DevDroid
CosFly-VLA: A Spatially Aware Vision-Language-Action Model for UAV Tracking28787ecd8dca672f
声明：CosFly-VLA: A Spatially Aware Vision-Language-Action Model for UAV Tracking
CosFly-VLA: A Spatially Aware Vision-Language-Action Model for UAV Tracking
The Reindustrial Revolution: Partnering with ABB Robotics, Universal Robots, and NVIDIA294664e0eeb98d78
声明：We have been working closely with NVIDIA to enable this pretraining at scale through NVIDIA Isaac Lab and Isaac Sim to create physically accurate simulations and NVIDIA Cosmos to generate and augment synthetic data, improving robustness and sim-to-real transfer.
We have been working closely with NVIDIA to enable this pretraining at scale through NVIDIA Isaac Lab and Isaac Sim to create physically accurate simulations and NVIDIA Cosmos to generate and augment synthetic data, improving robustness and sim-to-real transfer.
全球首个！银河通用新框架仅需人类视频即可部署，特斯拉蚌埠住了2a1206af49536cdb
声明：第一组是WAM-COTRAIN，团队直接把人类演示数据混进联合训练，结果显示任务成功率只有29.8%，反而低于完全不用人类数据的基线，这也说明了人类视频并不能简单作为新增训练数据加入预训练流程。
第一组是WAM-COTRAIN，团队直接把人类演示数据混进联合训练，结果显示任务成功率只有29.8%，反而低于完全不用人类数据的基线，这也说明了人类视频并不能简单作为新增训练数据加入预训练流程。
isaac-sim/IsaacLab v3.0.0-beta22aad648f9ebb32a8
声明：isaac-sim/IsaacLab v3.0.0-beta2
isaac-sim/IsaacLab v3.0.0-beta2
《人形机器人创新发展指导意见》解读2bb594451216baf8
声明：《人形机器人创新发展指导意见》解读
《人形机器人创新发展指导意见》解读
RoboTTT: Context Scaling for Robot Policies2c91dadc7595e310
声明：We introduce Test-Time-Training Robot Policies (RoboTTT), a robot model and training recipe that scale visuomotor context to 8K timesteps, three orders of magnitude beyond state-of-the-art policies, without growing inference latency.
We introduce Test-Time-Training Robot Policies (RoboTTT), a robot model and training recipe that scale visuomotor context to 8K timesteps, three orders of magnitude beyond state-of-the-art policies, without growing inference latency.
Our First Generalist Policy2d7fcf7796baeb09
声明：Our First Generalist Policy
Our First Generalist Policy
千寻智能完成5.28亿元Pre-A轮融资，领先VLA模型加速具身智能落地2f251b3037545679
声明：千寻智能（Spirit AI）成立于2024年2月，是国内领先的AI+机器人全栈生产力级技术能力的具身智能公司。公司团队成员来自UC Berkeley、CMU、清华大学、北京大学等顶尖高校，以及字节、小米、腾讯等知名企业。成立短短一年多，千寻智能凭借卓越的技术实力和创新理念，已成功完成多轮融资，并致力于打造世界级通用人形机器人和下一代具身大模型与学习算法，让通用的机器人伙伴走进千家万户，驱动世界迈向智能机器人时代。
千寻智能（Spirit AI）成立于2024年2月，是国内领先的AI+机器人全栈生产力级技术能力的具身智能公司。公司团队成员来自UC Berkeley、CMU、清华大学、北京大学等顶尖高校，以及字节、小米、腾讯等知名企业。成立短短一年多，千寻智能凭借卓越的技术实力和创新理念，已成功完成多轮融资，并致力于打造世界级通用人形机器人和下一代具身大模型与学习算法，让通用的机器人伙伴走进千家万户，驱动世界迈向智能机器人时代。
GEAR3173820a3e7dbe8a
声明：At GEAR, our mission is to build foundation models for embodied agents in virtual and physical worlds.
At GEAR, our mission is to build foundation models for embodied agents in virtual and physical worlds.
Qwen - Alibaba Cloud32ab89b344245512
声明：Multimodal models, including Qwen-VL (large language vision model), Qwen-TTS (text-to-speech model), and Qwen-Audio, support cross-modal processing, while the latest Qwen3-Omni is capable of processing and reasoning across text, audio, and vision modalities simultaneously.
Multimodal models, including Qwen-VL (large language vision model), Qwen-TTS (text-to-speech model), and Qwen-Audio, support cross-modal processing, while the latest Qwen3-Omni is capable of processing and reasoning across text, audio, and vision modalities simultaneously.
Hugging Face model lerobot/lingbot_va_libero_long3b3a2467c6fa9a08
声明：Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
NVIDIA/Isaac-GR00T n1.6.1-release3c1e8051215a272e
声明：Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_6/
Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_6/
isaac-sim/IsaacLab v3.0.0-beta3d98d1da47cbabed
声明：The previous `write_*_to_sim(data, env_ids)` pattern is replaced by: - `write_*_to_sim_index(data, env_ids)` — sparse indexed data - `write_*_to_sim_mask(data, env_mask)` — full data with boolean mask ### Deprecated APIs - `root_physx_view` → use `root_view` - `object_*` naming on `RigidObjectCollection` → use `body_*` - `--headless` CLI argument → use `--visualizer` / `--viz` - OpenXRDevice / OpenXRDeviceCfg → use IsaacTeleopDevice / IsaacTeleopCfg - Sensor `pose_w`, `pos_w`, `quat_w` properties → use `FrameTransformer` ### URDF & MJCF Importers Updated Both importers have been rewritten for Isaac Sim 6.0.
The previous `write_*_to_sim(data, env_ids)` pattern is replaced by: - `write_*_to_sim_index(data, env_ids)` — sparse indexed data - `write_*_to_sim_mask(data, env_mask)` — full data with boolean mask ### Deprecated APIs - `root_physx_view` → use `root_view` - `object_*` naming o…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
Hugging Face dataset agibot-world/GenieSimAssets4040806377ee4191
声明：Tags: task_categories:other, language:en, arxiv:2601.02078, region:us, real-world, dual-arm, robotics manipulation, simulation
Tags: task_categories:other, language:en, arxiv:2601.02078, region:us, real-world, dual-arm, robotics manipulation, simulation
Our First Generalist Policy41563d4f294ace4c
声明：Our first step is π0, a prototype model that combines large-scale multi-task and multi-robot data collection with a new network architecture to enable the most capable and dexterous generalist robot policy to date.
Our first step is π0, a prototype model that combines large-scale multi-task and multi-robot data collection with a new network architecture to enable the most capable and dexterous generalist robot policy to date.
How a Spinning Drone Exploits Your Eyes to Become Nearly Invisible47807e6af3f5b64a
声明：For a deeper dive into all the particulars of this project, read the paper: Computational Design of a Low-Visibility UAV Using a Human-Aligned Perceptual Metric, by Jingxian Wang, Chen Yu, David Matthews, Emma Alexander, Sam Kriegman, and Michael Rubenstein from Northwestern University, which is being presented this week at RSS 2026 in Sydney.
For a deeper dive into all the particulars of this project, read the paper: Computational Design of a Low-Visibility UAV Using a Human-Aligned Perceptual Metric, by Jingxian Wang, Chen Yu, David Matthews, Emma Alexander, Sam Kriegman, and Michael Rubenstein from Northwestern Univ…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
《人形机器人创新发展指导意见》解读4a1b5d56831349c5
声明：《指导意见》从特种领域、制造业典型场景、民生及重点行业三类方向提出意见措施。一是加快人形机器人在特种环境应用，面向恶劣条件、危险场景作业等需求，强化复杂环境下本体控制、快速移动、精确感知等能力。二是聚焦3C、汽车等制造业重点领域，提升人形机器人工具操作与任务执行能力，打造人形机器人示范产线和工厂，在典型制造场景实现深度应用。三是拓展人形机器人在医疗、家政等民生领域服务应用，满足生命健康、陪伴护理等高品质生活需求，推动人形机器人在农业、物流等重点行业应用落地，提升人机交互、灵巧抓取、分拣搬运、智能配送等作业能力。
《指导意见》从特种领域、制造业典型场景、民生及重点行业三类方向提出意见措施。一是加快人形机器人在特种环境应用，面向恶劣条件、危险场景作业等需求，强化复杂环境下本体控制、快速移动、精确感知等能力。二是聚焦3C、汽车等制造业重点领域，提升人形机器人工具操作与任务执行能力，打造人形机器人示范产线和工厂，在典型制造场景实现深度应用。三是拓展人形机器人在医疗、家政等民生领域服务应用，满足生命健康、陪伴护理等高品质生活需求，推动人形机器人在农业、物流等重点行业应用落地，提升人机交互、灵巧抓取、分拣搬运、智能配送等作业能力。
98年哈工大教授创业，要做人形灵巧操作世界模型4a1d0fad81686728
声明：NVIDIA EgoScale证明，第一人称人类操作数据在灵巧操作训练中具备Scaling Law；Generalist / Gen-1则进一步显示，大规模人类数据预训练结合少量真机数据后训练，可以显著提升机器人在长程、复杂任务中的成功率。
NVIDIA EgoScale证明，第一人称人类操作数据在灵巧操作训练中具备Scaling Law；Generalist / Gen-1则进一步显示，大规模人类数据预训练结合少量真机数据后训练，可以显著提升机器人在长程、复杂任务中的成功率。
GEAR4befca44c8f6f06a
声明：At GEAR, our mission is to build foundation models for embodied agents in virtual and physical worlds.
At GEAR, our mission is to build foundation models for embodied agents in virtual and physical worlds.
Hugging Face model lerobot/lingbot_va_libero_long4c966a9e5704d4a9
声明：Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
Hugging Face model lerobot/pi052_robocasa4dde42edc595abc0
声明：Tags: lerobot, safetensors, robotics, pi052, dataset:pepijn223/robocasa_pretrain_human300_v4_annotated5, license:apache-2.0, region:us
Tags: lerobot, safetensors, robotics, pi052, dataset:pepijn223/robocasa_pretrain_human300_v4_annotated5, license:apache-2.0, region:us
isaac-sim/IsaacLab v3.0.0-beta24f5f9ed9a3ad824d
声明：--- ## ✨ Highlights ### Multi-Backend Stabilization The multi-backend architecture introduced in the first 3.0 beta has been hardened across PhysX, Newton, OVPhysX, Isaac RTX, OVRTX, and kit-less execution paths.
--- ## ✨ Highlights ### Multi-Backend Stabilization The multi-backend architecture introduced in the first 3.0 beta has been hardened across PhysX, Newton, OVPhysX, Isaac RTX, OVRTX, and kit-less execution paths.
Scaling Behavior Foundation Model for Humanoid Robots52cdfb2119a53386
声明：Through extensive experiments in both simulation and real-world deployment, we demonstrate that our approach yields significant improvements in control fidelity and task generalization, reducing Mean Per-Keypoint Position Error (MPKPE) on the test set by over 10% in local mode and 82% in global mode compared with existing humanoid controllers.
Through extensive experiments in both simulation and real-world deployment, we demonstrate that our approach yields significant improvements in control fidelity and task generalization, reducing Mean Per-Keypoint Position Error (MPKPE) on the test set by over 10% in local mode an…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
《人形机器人创新发展指导意见》解读5382b6de6deefd76
声明：《指导意见》从特种领域、制造业典型场景、民生及重点行业三类方向提出意见措施。一是加快人形机器人在特种环境应用，面向恶劣条件、危险场景作业等需求，强化复杂环境下本体控制、快速移动、精确感知等能力。二是聚焦3C、汽车等制造业重点领域，提升人形机器人工具操作与任务执行能力，打造人形机器人示范产线和工厂，在典型制造场景实现深度应用。三是拓展人形机器人在医疗、家政等民生领域服务应用，满足生命健康、陪伴护理等高品质生活需求，推动人形机器人在农业、物流等重点行业应用落地，提升人机交互、灵巧抓取、分拣搬运、智能配送等作业能力。
《指导意见》从特种领域、制造业典型场景、民生及重点行业三类方向提出意见措施。一是加快人形机器人在特种环境应用，面向恶劣条件、危险场景作业等需求，强化复杂环境下本体控制、快速移动、精确感知等能力。二是聚焦3C、汽车等制造业重点领域，提升人形机器人工具操作与任务执行能力，打造人形机器人示范产线和工厂，在典型制造场景实现深度应用。三是拓展人形机器人在医疗、家政等民生领域服务应用，满足生命健康、陪伴护理等高品质生活需求，推动人形机器人在农业、物流等重点行业应用落地，提升人机交互、灵巧抓取、分拣搬运、智能配送等作业能力。
全球首个！银河通用新框架仅需人类视频即可部署，特斯拉蚌埠住了53b4eff7d4890144
声明：过去大家相信Scaling Law，模型能力主要来自更大的参数、更多的数据和更长时间的训练，但从o1这些推理模型开始，新的共识是模型能力来自推理阶段。
过去大家相信Scaling Law，模型能力主要来自更大的参数、更多的数据和更长时间的训练，但从o1这些推理模型开始，新的共识是模型能力来自推理阶段。
Hugging Face model lerobot/pi052_robocasa56519d5db1cd3bff
声明：Tags: lerobot, safetensors, robotics, pi052, dataset:pepijn223/robocasa_pretrain_human300_v4_annotated5, license:apache-2.0, region:us
Tags: lerobot, safetensors, robotics, pi052, dataset:pepijn223/robocasa_pretrain_human300_v4_annotated5, license:apache-2.0, region:us
Building the general-purpose robotic brain570c5a7dca6703b5
声明：Building the general-purpose robotic brain
Building the general-purpose robotic brain
Hugging Face dataset agibot-world/GenieSimAssets5817fe23c6ec1cb7
声明：Tags: task_categories:other, language:en, arxiv:2601.02078, region:us, real-world, dual-arm, robotics manipulation, simulation
Tags: task_categories:other, language:en, arxiv:2601.02078, region:us, real-world, dual-arm, robotics manipulation, simulation
GEAR5987b078c1b337bd
声明：We're looking for passionate researchers and engineers who want to work on cutting-edge problems in robotics, foundation models, and simulation.
We're looking for passionate researchers and engineers who want to work on cutting-edge problems in robotics, foundation models, and simulation.
千寻智能spirit AI5ad1a0445eb5f600
声明：千寻智能spirit AI
千寻智能spirit AI
Hugging Face dataset agibot-world/AgiBotWorldChallenge-20265b35e0220dfd7657
声明：Hugging Face dataset agibot-world/AgiBotWorldChallenge-2026
Hugging Face dataset agibot-world/AgiBotWorldChallenge-2026
Hugging Face dataset agibot-world/AgiBotWorld20265bca1e315481303e
声明：Hugging Face dataset agibot-world/AgiBotWorld2026
Hugging Face dataset agibot-world/AgiBotWorld2026
Hugging Face dataset agibot-world/GenieSim3.0-Dataset5e941821f5b248fe
声明：Hugging Face dataset agibot-world/GenieSim3.0-Dataset
Hugging Face dataset agibot-world/GenieSim3.0-Dataset
Qwen - Alibaba Cloud5f0899c2cf23e6a0
声明：Qwen - Alibaba Cloud
Qwen - Alibaba Cloud
How Toshio Fukuda Helped Shape IEEE and Global Robotics6160fc3e419de309
声明：After earning his doctorate, he returned to Japan in 1977 to work as a research scientist at the government’s Mechanical Engineering Laboratory, later renamed the National Institute of Advanced Industrial Science and Technology, in Tsukuba.
After earning his doctorate, he returned to Japan in 1977 to work as a research scientist at the government’s Mechanical Engineering Laboratory, later renamed the National Institute of Advanced Industrial Science and Technology, in Tsukuba.
近百名玩家涌入具身数据：一年融资44.7亿，谁能真靠“卖数据”赚钱？61e1c0fbd9d9962e
声明：近百名玩家涌入具身数据：一年融资44.7亿，谁能真靠“卖数据”赚钱？
近百名玩家涌入具身数据：一年融资44.7亿，谁能真靠“卖数据”赚钱？
Hugging Face model lerobot/lingbot_va_base62267347191a411a
声明：Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
How Cheap Ground Robots Are Rewriting Frontline Warfare in Ukraine65f48f1df4bc1eb1
声明：Oleg Fedoryshyn, the head of research and design at DevDroid, another prominent Ukrainian UGV developer, said the ground robots can be controlled from as far as 100 kilometers away using Starlink connectivity, LTE networks, or mesh-networked military radio systems.
Oleg Fedoryshyn, the head of research and design at DevDroid, another prominent Ukrainian UGV developer, said the ground robots can be controlled from as far as 100 kilometers away using Starlink connectivity, LTE networks, or mesh-networked military radio systems.
98年哈工大教授创业，要做人形灵巧操作世界模型65f943a754df2ff5
声明：98年哈工大教授创业，要做人形灵巧操作世界模型
98年哈工大教授创业，要做人形灵巧操作世界模型
huggingface/lerobot Release v0.5.068073ab88a87069b
声明：by @s1lent4gnt in https://github.com/huggingface/lerobot/pull/2482 * chore(deps): bump ceil datasets by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/2946 * feat(dataset): add streaming video encoding + HW encoder support by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/2974 * Add metadata_buffer_size to dataset creation by @yuecideng in https://github.com/huggingface/lerobot/pull/2998 * feat(motors): add RobStride CAN implementation by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/2821 * chore(docs): update the document for Phone teleop to clarify how to use the examples by @ynaka81 in https://github.com/huggingface/lerobot/pull/2991 * fix/
by @s1lent4gnt in https://github.com/huggingface/lerobot/pull/2482 * chore(deps): bump ceil datasets by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/2946 * feat(dataset): add streaming video encoding + HW encoder support by @imstevenpmwork in https://github.com/…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
GEAR681877ef7e3349e4
声明：GEAR
GEAR
Video Friday: A World Cup for Robots68400e6709684aa4
声明：Video Friday: A World Cup for Robots
Video Friday: A World Cup for Robots
Publications68e17ac532e1c000
声明：Our research contributes to the fields of AI, robotics, and embodied intelligence
Our research contributes to the fields of AI, robotics, and embodied intelligence
isaac-sim/IsaacLab v3.0.0-beta69097f024776e2ee
声明：| Best For | |---------|---------------------|----------| | **Isaac RTX** | Yes | Full sensor fidelity, photorealistic rendering | | **OVRTX** | No (+ `isaaclab_ov[ovrtx]`) | Kit-less RTX pipeline | | **Newton Warp** | No (kit-less) | Fast training without RTX | ### Pluggable Visualizer System Isaac Lab 3.0 introduces a new **pluggable visualizer framework** (`isaaclab_visualizers`) with four interchangeable backends, all decoupled from the physics engine and renderer: | Visualizer | Best For | Key Features | |---|---|---| | **Omniverse (Kit)** | High-fidelity, Isaac Sim integration | USD stage, visual markers, live training plots | | **Newton** | Fast iteration, low overhead | OpenGL render
| Best For | |---------|---------------------|----------| | **Isaac RTX** | Yes | Full sensor fidelity, photorealistic rendering | | **OVRTX** | No (+ `isaaclab_ov[ovrtx]`) | Kit-less RTX pipeline | | **Newton Warp** | No (kit-less) | Fast training without RTX | ### Pluggable Vis…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
huggingface/lerobot Release v0.5.06c135bef695c4040
声明：## Release Blog https://huggingface.co/blog/lerobot-release-v050 ## What's Changed * chore(dependencies): Bump lerobot to 0.4.5 by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/3051 * chore: add AI policy by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/3055 * Improve policy_device documentation for async.mdx by @btelles in https://github.com/huggingface/lerobot/pull/3060 * fix(frame_index): making rerun's "frame_index" timeline compatible with behaviour1k datasets by @CarolinePascal in https://github.com/huggingface/lerobot/pull/3068 * fix(dataset edit tools): clarifying `root` argument usage + adding related features by @CarolinePascal in https://github.co
## Release Blog https://huggingface.co/blog/lerobot-release-v050 ## What's Changed * chore(dependencies): Bump lerobot to 0.4.5 by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/3051 * chore: add AI policy by @imstevenpmwork in https://github.com/huggingface/lerob…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
NVIDIA/Isaac-GR00T n1.5 release6c168684e54ef2b9
声明：Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_5/
Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_5/
Qwen - Alibaba Cloud6c8286ccaa440bbd
声明：Qwen - Alibaba Cloud
Qwen - Alibaba Cloud
手握英伟达宁德时代相同剧本，具身智能的第一个「基建商」出现了6de196af235c1798
声明：这套逻辑有现成的参照系——2017年的英伟达，当时更多被看作游戏显卡供应商，没人预料到它会成为AI浪潮中最绕不开的存在。那一年行业都在追算法、追工具，算力被视为“供应链的一环”而非战略制高点。
这套逻辑有现成的参照系——2017年的英伟达，当时更多被看作游戏显卡供应商，没人预料到它会成为AI浪潮中最绕不开的存在。那一年行业都在追算法、追工具，算力被视为“供应链的一环”而非战略制高点。
How a Spinning Drone Exploits Your Eyes to Become Nearly Invisible6e9197a2f9871a9e
声明：The bits that you need for this kind of drone include the motor and propeller, a couple of batteries, a controller, some counterweights (which could be replaced with more batteries or payload), 0.8mm carbon fiber rods to tie it all together, and a connector for the handheld launcher that gets the whole thing up to speed.
The bits that you need for this kind of drone include the motor and propeller, a couple of batteries, a controller, some counterweights (which could be replaced with more batteries or payload), 0.8mm carbon fiber rods to tie it all together, and a connector for the handheld launc…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
Qwen - Alibaba Cloud709d2d9f02eb2db3
声明：The latest Qwen3 models adopt hybrid thinking modes ("Thinking" and "Non-Thinking") that allow you to flexibly control reasoning performance, speed, and costs.
The latest Qwen3 models adopt hybrid thinking modes ("Thinking" and "Non-Thinking") that allow you to flexibly control reasoning performance, speed, and costs.
DriftWorld: Fast World Modeling through Drifting70ef99ab78b44df8
声明：Rather than denoising iteratively at inference, DriftWorld learns an action-conditioned drift during training, allowing it to generate future frames from the current observation and a candidate action sequence in a single forward pass at 30+ fps, which is 17x faster on average than diffusion based baselines.
Rather than denoising iteratively at inference, DriftWorld learns an action-conditioned drift during training, allowing it to generate future frames from the current observation and a candidate action sequence in a single forward pass at 30+ fps, which is 17x faster on average th…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
创造人机竞速新历史——人形机器人马拉松见证中国未来产业奔跑之路72a5654dab044ed2
声明：创造人机竞速新历史——人形机器人马拉松见证中国未来产业奔跑之路
创造人机竞速新历史——人形机器人马拉松见证中国未来产业奔跑之路
千寻智能Spirit AI741d8d9a19e13355
声明：凭借 AI+机器人全栈生产力级技术能力，千寻智能荣获2024 中国科技机器人企业TOP50、2024年度最具投资价值企业TOP100、LeadeRobot2024年度机器人科学引领奖/清华、leaderobot2024年度具身智能卓越性能奖、机器人行业最具投资价值企业 TOP 50、2024科技产业最具商业潜力榜、2024人工智能最具价值创业公司TOP20、2024 人工智能年度焦点人物、2024创业邦 100 未来独角兽、极新 AGI独角兽 top50、ISIG 产业大会-最佳技术创新奖、CSDN2024中国开发者影响力年度评选-人工智能企业、人形机器人商业落地 top20、具身智能算法革新奖、最具投资价值 top50、2025年值得关注的AIGC企业、2025机器人产业领军企业 TOP100、2025 杭州市准独角兽、2025 年中国未来独角兽、创业之星、优势成长领军企业等
凭借 AI+机器人全栈生产力级技术能力，千寻智能荣获2024 中国科技机器人企业TOP50、2024年度最具投资价值企业TOP100、LeadeRobot2024年度机器人科学引领奖/清华、leaderobot2024年度具身智能卓越性能奖、机器人行业最具投资价值企业 TOP 50、2024科技产业最具商业潜力榜、2024人工智能最具价值创业公司TOP20、2024 人工智能年度焦点人物、2024创业邦 100 未来独角兽、极新 AGI独角兽 top50、ISIG 产业大会-最佳技术创新奖、CSDN2024中国开发者影响力年度评选-人工智能企业、人形机器…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
huggingface/lerobot Release v0.5.07602fb64851e6857
声明：huggingface/lerobot Release v0.5.0
huggingface/lerobot Release v0.5.0
The Reindustrial Revolution: Partnering with ABB Robotics, Universal Robots, and NVIDIA766c353723f9f171
声明：This is, however, a catch-22 : robots need data to improve, but only capable robots can be deployed to gather it.
This is, however, a catch-22 : robots need data to improve, but only capable robots can be deployed to gather it.
Home | Robotics: Science and Systems7dbc3d777c1e3263
声明：We are pleased to announce the 22nd edition of the “Robotics: Science and Systems” (RSS) conference to be held at the University of Technology Sydney and the International Convention Centre, Sydney, Australia from July 13-17, 2026.
We are pleased to announce the 22nd edition of the “Robotics: Science and Systems” (RSS) conference to be held at the University of Technology Sydney and the International Convention Centre, Sydney, Australia from July 13-17, 2026.
创造人机竞速新历史——人形机器人马拉松见证中国未来产业奔跑之路7ffd2a132e9592c1
声明：场景落地是创新成果转化的关键载体，这一点在马拉松赛事中体现得尤为直观。“一方面通过大众最为熟悉的马拉松比赛，吸引全民关注，为产业发展营造良好氛围，另一方面，在20多公里的复杂赛道上，面向真实场景的‘极限压力测试’，让机器人持续稳定运行、应对各类路况，引导企业攻克核心技术难题。”中国电子学会副秘书长梁靓说。
场景落地是创新成果转化的关键载体，这一点在马拉松赛事中体现得尤为直观。“一方面通过大众最为熟悉的马拉松比赛，吸引全民关注，为产业发展营造良好氛围，另一方面，在20多公里的复杂赛道上，面向真实场景的‘极限压力测试’，让机器人持续稳定运行、应对各类路况，引导企业攻克核心技术难题。”中国电子学会副秘书长梁靓说。
手握英伟达宁德时代相同剧本，具身智能的第一个「基建商」出现了805360245f73e799
声明：手握英伟达宁德时代相同剧本，具身智能的第一个「基建商」出现了
手握英伟达宁德时代相同剧本，具身智能的第一个「基建商」出现了
NVIDIA/Isaac-GR00T n1.6.1-release806c9bf14ed7dd32
声明：Early Access n1.7 release
Early Access n1.7 release
Home | Robotics: Science and Systems832742455e012937
声明：Home | Robotics: Science and Systems
Home | Robotics: Science and Systems
huggingface/lerobot Release v0.5.083f7d224adb74b4e
声明：huggingface/lerobot Release v0.4.4
huggingface/lerobot Release v0.4.4
千寻智能Spirit AI84eca31e10753ff5
声明：千寻智能Spirit AI
千寻智能Spirit AI
《人形机器人创新发展指导意见》解读86dadf0d05f09eb7
声明：《人形机器人创新发展指导意见》解读
《人形机器人创新发展指导意见》解读
Scaling Behavior Foundation Model for Humanoid Robots8be6dcdbbcb080cd
声明：Scaling Behavior Foundation Model for Humanoid Robots
Scaling Behavior Foundation Model for Humanoid Robots
Our First Generalist Policy8cdea5b57577f82a
声明：Over the past eight months, we’ve developed a general-purpose robot foundation model that we call π0 (pi-zero).
Over the past eight months, we’ve developed a general-purpose robot foundation model that we call π0 (pi-zero).
isaac-sim/IsaacLab v3.0.0-beta8e8da3630284c63c
声明：isaac-sim/IsaacLab v3.0.0-beta
isaac-sim/IsaacLab v3.0.0-beta
《人形机器人创新发展指导意见》解读8f6209771ec3611b
声明：《指导意见》部署了5方面任务：在关键技术突破方面，打造人形机器人“大脑”和“小脑”、突破“肢体”关键技术、健全技术创新体系。在产品培育方面，打造整机产品、夯实基础部组件、推动软件创新。在场景拓展方面，服务特种领域需求、打造制造业典型场景、加快民生及重点行业推广。在生态营造方面，培育优质企业、完善创新载体和开源环境、推动产业集聚发展。在支撑能力方面，健全产业标准体系、提升检验检测和中试验证能力、加强安全治理能力。
《指导意见》部署了5方面任务：在关键技术突破方面，打造人形机器人“大脑”和“小脑”、突破“肢体”关键技术、健全技术创新体系。在产品培育方面，打造整机产品、夯实基础部组件、推动软件创新。在场景拓展方面，服务特种领域需求、打造制造业典型场景、加快民生及重点行业推广。在生态营造方面，培育优质企业、完善创新载体和开源环境、推动产业集聚发展。在支撑能力方面，健全产业标准体系、提升检验检测和中试验证能力、加强安全治理能力。
A Steerable Model with Emergent Capabilities8ffcebcdcc99c25d
声明：These episodes look quite different from what the mobile robot actually does in our experiments, suggesting that π0.7 can generalize and compose behaviors to load the sweet potato into the air fryer, much like how an LLM composes different parts of text seen in large-scale datasets from the web:
These episodes look quite different from what the mobile robot actually does in our experiments, suggesting that π0.7 can generalize and compose behaviors to load the sweet potato into the air fryer, much like how an LLM composes different parts of text seen in large-scale datase…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
98年哈工大教授创业，要做人形灵巧操作世界模型9a461c1c073c17db
声明：公开信息显示，他26岁获评哈工大（深圳）计算机学院长聘教授、博导，曾获Google Ph.D.
公开信息显示，他26岁获评哈工大（深圳）计算机学院长聘教授、博导，曾获Google Ph.D.
《人形机器人创新发展指导意见》解读9bbe51d51e28bb0e
声明：《指导意见》部署了5方面任务：在关键技术突破方面，打造人形机器人“大脑”和“小脑”、突破“肢体”关键技术、健全技术创新体系。在产品培育方面，打造整机产品、夯实基础部组件、推动软件创新。在场景拓展方面，服务特种领域需求、打造制造业典型场景、加快民生及重点行业推广。在生态营造方面，培育优质企业、完善创新载体和开源环境、推动产业集聚发展。在支撑能力方面，健全产业标准体系、提升检验检测和中试验证能力、加强安全治理能力。
《指导意见》部署了5方面任务：在关键技术突破方面，打造人形机器人“大脑”和“小脑”、突破“肢体”关键技术、健全技术创新体系。在产品培育方面，打造整机产品、夯实基础部组件、推动软件创新。在场景拓展方面，服务特种领域需求、打造制造业典型场景、加快民生及重点行业推广。在生态营造方面，培育优质企业、完善创新载体和开源环境、推动产业集聚发展。在支撑能力方面，健全产业标准体系、提升检验检测和中试验证能力、加强安全治理能力。
千寻智能完成5.28亿元Pre-A轮融资，领先VLA模型加速具身智能落地9bca78116b7668a9
声明：千寻智能完成5.28亿元Pre-A轮融资，领先VLA模型加速具身智能落地
千寻智能完成5.28亿元Pre-A轮融资，领先VLA模型加速具身智能落地
isaac-sim/IsaacLab v3.0.0-beta2.patch19c07fac4b1d0b5c7
声明：isaac-sim/IsaacLab v3.0.0-beta2.patch1
isaac-sim/IsaacLab v3.0.0-beta2.patch1
Publications9c88aacda30193ee
声明：Publications
Publications
Qwen - Alibaba Cloud9f53191720f4b66d
声明：The latest Qwen3 models adopt hybrid thinking modes ("Thinking" and "Non-Thinking") that allow you to flexibly control reasoning performance, speed, and costs.
The latest Qwen3 models adopt hybrid thinking modes ("Thinking" and "Non-Thinking") that allow you to flexibly control reasoning performance, speed, and costs.
Moravec's Paradox and the Robot Olympics9f783be2f4a1bff1
声明：Moravec's Paradox and the Robot Olympics
Moravec's Paradox and the Robot Olympics
Hugging Face dataset agibot-world/GenieSimAssets9fc708b2a069aa2d
声明：Hugging Face dataset agibot-world/GenieSimAssets
Hugging Face dataset agibot-world/GenieSimAssets
千寻智能完成5.28亿元Pre-A轮融资，领先VLA模型加速具身智能落地a046db13e94571b5
声明：本轮领投方Prosperity7 Ventures(P7)执行董事总经理 Raed Twaily表示：“千寻智能在 VLA 泛化能力和模型迭代速度方面所展现出的实力，让我们很关注。其 VLA 的泛化能力，赋予了机器人适应复杂多变环境与多样化任务的本领，使其能够精准执行连续复杂操作，我们对千寻智能基于国际前沿理念构建的技术路径予以高度认可，放眼未来，期待千寻智能持续发力，不断拓展具身智能的技术边界，在全球舞台上引领具身智能技术发展新潮流，为全球智能机器人产业变革注入强劲动力。”
本轮领投方Prosperity7 Ventures(P7)执行董事总经理 Raed Twaily表示：“千寻智能在 VLA 泛化能力和模型迭代速度方面所展现出的实力，让我们很关注。其 VLA 的泛化能力，赋予了机器人适应复杂多变环境与多样化任务的本领，使其能够精准执行连续复杂操作，我们对千寻智能基于国际前沿理念构建的技术路径予以高度认可，放眼未来，期待千寻智能持续发力，不断拓展具身智能的技术边界，在全球舞台上引领具身智能技术发展新潮流，为全球智能机器人产业变革注入强劲动力。”
huggingface/lerobot Release v0.6.0a1831735cc56344b
声明：## Release Blog https://huggingface.co/blog/lerobot-release-v060 ## ⚠️ Breaking changes (v0.5.1 -> v0.6.0) v0.6.0 cleans house, and a few changes need your attention when upgrading: - `pip install lerobot` no longer includes dataset or training dependencies; add the extra you need (e.g.
## Release Blog https://huggingface.co/blog/lerobot-release-v060 ## ⚠️ Breaking changes (v0.5.1 -> v0.6.0) v0.6.0 cleans house, and a few changes need your attention when upgrading: - `pip install lerobot` no longer includes dataset or training dependencies; add the extra you nee…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
How a Spinning Drone Exploits Your Eyes to Become Nearly Invisiblea2a9de9ef0612696
声明：How a Spinning Drone Exploits Your Eyes to Become Nearly Invisible
How a Spinning Drone Exploits Your Eyes to Become Nearly Invisible
The Reindustrial Revolution: Partnering with ABB Robotics, Universal Robots, and NVIDIAa2cfe872162b7d5a
声明：The Reindustrial Revolution: Partnering with ABB Robotics, Universal Robots, and NVIDIA
The Reindustrial Revolution: Partnering with ABB Robotics, Universal Robots, and NVIDIA
NVIDIA/Isaac-GR00T n1.5 releasea321a00236d22e6a
声明：NVIDIA/Isaac-GR00T n1.5 release
NVIDIA/Isaac-GR00T n1.5 release
千寻智能完成5.28亿元Pre-A轮融资，领先VLA模型加速具身智能落地a3616b6d0d81a0f2
声明：本轮领投方Prosperity7 Ventures(P7)执行董事总经理 Raed Twaily表示：“千寻智能在 VLA 泛化能力和模型迭代速度方面所展现出的实力，让我们很关注。其 VLA 的泛化能力，赋予了机器人适应复杂多变环境与多样化任务的本领，使其能够精准执行连续复杂操作，我们对千寻智能基于国际前沿理念构建的技术路径予以高度认可，放眼未来，期待千寻智能持续发力，不断拓展具身智能的技术边界，在全球舞台上引领具身智能技术发展新潮流，为全球智能机器人产业变革注入强劲动力。”
本轮领投方Prosperity7 Ventures(P7)执行董事总经理 Raed Twaily表示：“千寻智能在 VLA 泛化能力和模型迭代速度方面所展现出的实力，让我们很关注。其 VLA 的泛化能力，赋予了机器人适应复杂多变环境与多样化任务的本领，使其能够精准执行连续复杂操作，我们对千寻智能基于国际前沿理念构建的技术路径予以高度认可，放眼未来，期待千寻智能持续发力，不断拓展具身智能的技术边界，在全球舞台上引领具身智能技术发展新潮流，为全球智能机器人产业变革注入强劲动力。”
Hugging Face model lerobot/lingbot_va_basea48c385fc23ea94f
声明：Hugging Face model lerobot/lingbot_va_base
Hugging Face model lerobot/lingbot_va_base
全球首个！银河通用新框架仅需人类视频即可部署，特斯拉蚌埠住了a5d9ebf915da6422
声明：全球首个！银河通用新框架仅需人类视频即可部署，特斯拉蚌埠住了
全球首个！银河通用新框架仅需人类视频即可部署，特斯拉蚌埠住了
AeroAct: Action-Centered World-Action Models for Language-Conditioned Quadrotor Flighta6a6cd2bfbbd411d
声明：AeroAct: Action-Centered World-Action Models for Language-Conditioned Quadrotor Flight
AeroAct: Action-Centered World-Action Models for Language-Conditioned Quadrotor Flight
GEARa731ea348464ff41
声明：We're looking for passionate researchers and engineers who want to work on cutting-edge problems in robotics, foundation models, and simulation.
We're looking for passionate researchers and engineers who want to work on cutting-edge problems in robotics, foundation models, and simulation.
Moravec's Paradox and the Robot Olympicsa764a1d1f04fff0a
声明：Vision-language-action models like π0.6 provide a way to capture general physical knowledge from a highly diverse repertoire of tasks, providing a powerful foundation from which to learn downstream skills with much smaller and more practical datasets.
Vision-language-action models like π0.6 provide a way to capture general physical knowledge from a highly diverse repertoire of tasks, providing a powerful foundation from which to learn downstream skills with much smaller and more practical datasets.
huggingface/lerobot Release v0.5.1aa2c3d4436a0f86f
声明：## What's Changed * chore(dependencies): Bump lerobot to 0.5.1 by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/3118 * fix(unitree_g1): correct SDK detection and update installation docs by @nepyope in https://github.com/huggingface/lerobot/pull/3115 * docs: fix HF_USER to ignore bash ANSI codes by @j3soon in https://github.com/huggingface/lerobot/pull/3119 * chore(dependencies): Increase opencv-python-headless upper bound by @traversaro in https://github.com/huggingface/lerobot/pull/3120 * Fix action padding key at SmolVLA by @ymd-h in https://github.com/huggingface/lerobot/pull/1717 * Update Dockerfile.user by @heuzef in https://github.com/huggingface/lerobot/pull/1633 * f
## What's Changed * chore(dependencies): Bump lerobot to 0.5.1 by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/3118 * fix(unitree_g1): correct SDK detection and update installation docs by @nepyope in https://github.com/huggingface/lerobot/pull/3115 * docs: fix…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
isaac-sim/IsaacLab v3.0.0-beta2aafb251eb11b427f
声明：It builds on [v3.0.0-beta](https://github.com/isaac-sim/IsaacLab/releases/tag/v3.0.0-beta) with additional features and improvements on Newton support (VBD, solver coupling, Kamino, rough terrain, sensors), multi-backend physics, simplified training and installation commands, kit-less workflows, visualizers, rendering, teleoperation, learning exports, installation, CI, and documentation.
It builds on [v3.0.0-beta](https://github.com/isaac-sim/IsaacLab/releases/tag/v3.0.0-beta) with additional features and improvements on Newton support (VBD, solver coupling, Kamino, rough terrain, sensors), multi-backend physics, simplified training and installation commands, kit…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
Inside X Square Robot’s Bold Plan for Real Home Robotsab5d5640a9b10de0
声明：Inside X Square Robot’s Bold Plan for Real Home Robots
Inside X Square Robot’s Bold Plan for Real Home Robots
NVIDIA/Isaac-GR00T n1.6.1-releaseb04c3da99cc0f156
声明：NVIDIA/Isaac-GR00T n1.7 release
NVIDIA/Isaac-GR00T n1.7 release
Qwen - Alibaba Cloudb08e8ad4b1b6bfc1
声明：Multimodal models, including Qwen-VL (large language vision model), Qwen-TTS (text-to-speech model), and Qwen-Audio, support cross-modal processing, while the latest Qwen3-Omni is capable of processing and reasoning across text, audio, and vision modalities simultaneously.
Multimodal models, including Qwen-VL (large language vision model), Qwen-TTS (text-to-speech model), and Qwen-Audio, support cross-modal processing, while the latest Qwen3-Omni is capable of processing and reasoning across text, audio, and vision modalities simultaneously.
isaac-sim/IsaacLab v2.3.2b16e5ea16e897d77
声明：## ✨ New Features ### Core & Simulation * Adds Raycaster with tracking support for dynamic meshes by @renezurbruegg in #3298 * Adds visual-based tactile sensor with shape sensing example by @JuanaDd in #3420 * Adds wrench composers allowing the composition of multiple wrenches on the same bodies by @AntoineRichard in #3287 * Adds multirotor/thruster actuator, multirotor asset and manager-based ARL drone task #3760 by @mihirk284 @grzemal @Zwoelf12 * Adds automatic transform discovery for IMU sensors to find valid parent bodies by @bmccann-bdai in #3864 * Adds friction force reporting to ContactSensor by @gattra-rai in #3563 * Adds MJCF spawner for importing MJCF-based assets by @KyleM73 in #1
## ✨ New Features ### Core & Simulation * Adds Raycaster with tracking support for dynamic meshes by @renezurbruegg in #3298 * Adds visual-based tactile sensor with shape sensing example by @JuanaDd in #3420 * Adds wrench composers allowing the composition of multiple wrenches on…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
Building the general-purpose robotic brainb376ebff2f51f8fa
声明：Building the general-purpose robotic brain
Building the general-purpose robotic brain
DriftWorld: Fast World Modeling through Driftingb7cd2f56fba83e64
声明：Beyond online control, DriftWorld can also serve as an offline simulator for ranking real-world robot policies, with rollout-based scores correlating with ground truth at up to 0.99.
Beyond online control, DriftWorld can also serve as an offline simulator for ranking real-world robot policies, with rollout-based scores correlating with ground truth at up to 0.99.
Hugging Face dataset agibot-world/GenieSim3.0-Datasetb91d2c71cb16cbf5
声明：Official Hugging Face dataset update.
Official Hugging Face dataset update.
AeroAct: Action-Centered World-Action Models for Language-Conditioned Quadrotor Flightb9611b6246bffa38
声明：Language-conditioned quadrotor flight requires a policy to ground semantic goals, anticipate the visual consequences of ego-motion, and output control references that remain smooth and dynamically executable under rapidly changing first-person views.
Language-conditioned quadrotor flight requires a policy to ground semantic goals, anticipate the visual consequences of ego-motion, and output control references that remain smooth and dynamically executable under rapidly changing first-person views.
Hugging Face model lerobot/lingbot_va_robotwinbb5b1c5fc9530580
声明：Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
Qwen - Alibaba Cloudbc09de90ae8b06fc
声明：Qwen - Alibaba Cloud
Qwen - Alibaba Cloud
Qwen - Alibaba Cloudbc18540ff3ef3ec7
声明：Multimodal models, including Qwen-VL (large language vision model), Qwen-TTS (text-to-speech model), and Qwen-Audio, support cross-modal processing, while the latest Qwen3-Omni is capable of processing and reasoning across text, audio, and vision modalities simultaneously.
Multimodal models, including Qwen-VL (large language vision model), Qwen-TTS (text-to-speech model), and Qwen-Audio, support cross-modal processing, while the latest Qwen3-Omni is capable of processing and reasoning across text, audio, and vision modalities simultaneously.
千寻智能spirit AIbcc4080d518f9050
声明：千寻智能（Spirit AI）成立于2024年1月，是一家全球领先的专注于构建机器人“通用大脑”的具身智能公司。公司致力于研发通用具身大模型，赋予机器人跨场景的卓越泛化性与精细的物理交互能力，解决具身智能从虚拟到现实落地的核心难题。作为行业领军者，千寻智能自研的 Spirit 系列模型（如 Spirit v1.5）在 RoboChallenge 全球权威基准测试中持续领跑，代表了当前具身智能模型能力最领先的技术水平。千寻智能凭借卓越的技术实力和创新理念，致力于让通用的机器人伙伴走进千家万户，驱动世界迈向智能机器人时代。
千寻智能（Spirit AI）成立于2024年1月，是一家全球领先的专注于构建机器人“通用大脑”的具身智能公司。公司致力于研发通用具身大模型，赋予机器人跨场景的卓越泛化性与精细的物理交互能力，解决具身智能从虚拟到现实落地的核心难题。作为行业领军者，千寻智能自研的 Spirit 系列模型（如 Spirit v1.5）在 RoboChallenge 全球权威基准测试中持续领跑，代表了当前具身智能模型能力最领先的技术水平。千寻智能凭借卓越的技术实力和创新理念，致力于让通用的机器人伙伴走进千家万户，驱动世界迈向智能机器人时代。
huggingface/lerobot Release v0.6.0bd06a3cc710b4724
声明：https://github.com/user-attachments/assets/0d2938ff-ee35-4195-ae80-d7dacb50dc2f ## What's Changed * chore(dependencies): Bump lerobot to 0.5.2 by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/3307 * refactor(envs): move benchmark dispatch into EnvConfig subclasses by @pkooij in https://github.com/huggingface/lerobot/pull/3272 * feat(envs): lazy env init + AsyncVectorEnv as default for n_envs > 1 by @pkooij in https://github.com/huggingface/lerobot/pull/3274 * feat(ci): add agent assitance workflow by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/3332 * chore(security): update claude.yml by @hf-security-analysis[bot] in https://github.com/huggingface/lerobot/
https://github.com/user-attachments/assets/0d2938ff-ee35-4195-ae80-d7dacb50dc2f ## What's Changed * chore(dependencies): Bump lerobot to 0.5.2 by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/3307 * refactor(envs): move benchmark dispatch into EnvConfig subclasse…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
Building the general-purpose robotic brainbe046aa240e42d8f
声明：Thus, a lot of researchers and competitors side-step the problem: start with an existing vision-and-language model (VLM) and sprinkle in less than 1% of real-world robot data to build a “robotics foundation model”.
Thus, a lot of researchers and competitors side-step the problem: start with an existing vision-and-language model (VLM) and sprinkle in less than 1% of real-world robot data to build a “robotics foundation model”.
NVIDIA/Isaac-GR00T n1.6.1-releasebfe621685b91c5a0
声明：NVIDIA/Isaac-GR00T n1.6 release
NVIDIA/Isaac-GR00T n1.6 release
Hugging Face model lerobot/pi052_robocasac248799ac5c2503e
声明：Hugging Face model lerobot/pi052_robocasa
Hugging Face model lerobot/pi052_robocasa
Mediac26b67b72aa213a5
声明：Media
Media
Qwen - Alibaba Cloudc543c05d822c8f3e
声明：The latest Qwen3 models adopt hybrid thinking modes ("Thinking" and "Non-Thinking") that allow you to flexibly control reasoning performance, speed, and costs.
The latest Qwen3 models adopt hybrid thinking modes ("Thinking" and "Non-Thinking") that allow you to flexibly control reasoning performance, speed, and costs.
Qwen - Alibaba Cloudc79ea590d801a143
声明：Qwen - Alibaba Cloud
Qwen - Alibaba Cloud
用世界模型给VLA当教练，原力灵机发布DW0.5，把RL搬进虚拟世界cc7ea8a75845c377
声明：不光能做出机械臂正常作业的视频，就算参考错误动作，也能还原任务失败的场景，以此支撑DFOL2.0框架的在线强化学习训练。
不光能做出机械臂正常作业的视频，就算参考错误动作，也能还原任务失败的场景，以此支撑DFOL2.0框架的在线强化学习训练。
NVIDIA/Isaac-GR00T n1.6.1-releasece3c8db50d9caa12
声明：Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_6/
Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_6/
Building the general-purpose robotic braind32c188c5c763959
声明：Our foundation model, the Skild Brain, which works on various quadrupeds, humanoids, table-top arms, mobile manipulators and more, follows a hierarchical architecture: (1) a low-frequency high-level manipulation and navigation action policy which provides inputs to a (2) high-frequency low-level action policy.
Our foundation model, the Skild Brain, which works on various quadrupeds, humanoids, table-top arms, mobile manipulators and more, follows a hierarchical architecture: (1) a low-frequency high-level manipulation and navigation action policy which provides inputs to a (2) high-fre…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
Qwen - Alibaba Cloudd48717404cf67918
声明：Multimodal models, including Qwen-VL (large language vision model), Qwen-TTS (text-to-speech model), and Qwen-Audio, support cross-modal processing, while the latest Qwen3-Omni is capable of processing and reasoning across text, audio, and vision modalities simultaneously.
Multimodal models, including Qwen-VL (large language vision model), Qwen-TTS (text-to-speech model), and Qwen-Audio, support cross-modal processing, while the latest Qwen3-Omni is capable of processing and reasoning across text, audio, and vision modalities simultaneously.
huggingface/lerobot Release v0.6.0d4f7592cc5e2d2f4
声明：huggingface/lerobot Release v0.6.0
huggingface/lerobot Release v0.6.0
Hugging Face model lerobot/lingbot_va_robotwind57f8b546f81510e
声明：Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
Video Friday: A World Cup for Robotsd5e5c1383092b5c6
声明：GEN-1 unlocks commercial viability across a broad range of applications—and while it cannot solve all tasks today, it is a significant step toward our mission of creating generalist intelligence for the physical world.
GEN-1 unlocks commercial viability across a broad range of applications—and while it cannot solve all tasks today, it is a significant step toward our mission of creating generalist intelligence for the physical world.
Hugging Face model lerobot/lingbot_va_based64b7096866842c0
声明：Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
Tags: lerobot, safetensors, lingbot_va, robotics, license:apache-2.0, region:us
Video Friday: A World Cup for Robotsdb321715f15088b6
声明：It improves average success rates to 99 percent on tasks where previous models achieve 64 percent, completes tasks roughly 3x faster than state-of-the-art, and requires only one hour of robot data for each of these results.
It improves average success rates to 99 percent on tasks where previous models achieve 64 percent, completes tasks roughly 3x faster than state-of-the-art, and requires only one hour of robot data for each of these results.
《人形机器人创新发展指导意见》解读e28051de656bf46a
声明：《指导意见》部署了5方面任务：在关键技术突破方面，打造人形机器人“大脑”和“小脑”、突破“肢体”关键技术、健全技术创新体系。在产品培育方面，打造整机产品、夯实基础部组件、推动软件创新。在场景拓展方面，服务特种领域需求、打造制造业典型场景、加快民生及重点行业推广。在生态营造方面，培育优质企业、完善创新载体和开源环境、推动产业集聚发展。在支撑能力方面，健全产业标准体系、提升检验检测和中试验证能力、加强安全治理能力。
《指导意见》部署了5方面任务：在关键技术突破方面，打造人形机器人“大脑”和“小脑”、突破“肢体”关键技术、健全技术创新体系。在产品培育方面，打造整机产品、夯实基础部组件、推动软件创新。在场景拓展方面，服务特种领域需求、打造制造业典型场景、加快民生及重点行业推广。在生态营造方面，培育优质企业、完善创新载体和开源环境、推动产业集聚发展。在支撑能力方面，健全产业标准体系、提升检验检测和中试验证能力、加强安全治理能力。
Hugging Face model lerobot/lingbot_va_libero_longe290f1dcd7ad2b5a
声明：Hugging Face model lerobot/lingbot_va_libero_long
Hugging Face model lerobot/lingbot_va_libero_long
Hugging Face dataset agibot-world/AgiBotWorld2026e34a35839322b922
声明：Tags: task_categories:robotics, language:en, license:cc-by-nc-sa-4.0, size_categories:1K<n<10K, modality:image, modality:text, region:us, agibot, imitation-learning, embodied-ai, lerobot, real-world, dual-arm
Tags: task_categories:robotics, language:en, license:cc-by-nc-sa-4.0, size_categories:1K<n<10K, modality:image, modality:text, region:us, agibot, imitation-learning, embodied-ai, lerobot, real-world, dual-arm
NVIDIA/Isaac-GR00T n1.5 releasee55295c7141b1a6a
声明：Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_5/
Research Blog https://research.nvidia.com/labs/gear/gr00t-n1_5/
Mediae68ecfd88768886f
声明：From Neuralangelo's high-fidelity neural surface reconstruction to Magic3D's text-to-3D content creation, these projects push the boundaries of innovation in AI.
From Neuralangelo's high-fidelity neural surface reconstruction to Magic3D's text-to-3D content creation, these projects push the boundaries of innovation in AI.
千寻智能spirit AIe6ccc17bb99c8c93
声明：北京市海淀区西北旺东路10号院东区5号楼中关村软件园互联网创新中心A区 406
北京市海淀区西北旺东路10号院东区5号楼中关村软件园互联网创新中心A区 406
Mediae80405f762fbdede
声明：NVIDIA today announced Project GR00T, a general-purpose foundation model for humanoid robots, designed to further its work driving breakthroughs in robotics and embodied AI.
NVIDIA today announced Project GR00T, a general-purpose foundation model for humanoid robots, designed to further its work driving breakthroughs in robotics and embodied AI.
GEARe92b64a3c8feca03
声明：GEAR
GEAR
千寻智能Spirit AIe961e57e5b30141c
声明：2024年1月，千寻智能（杭州）科技有限公司正式注册成立。8月，获得近2亿元种子轮+天使轮融资，由弘晖基金领投，达晨创投、千乘资本跟投，同时，顺为资本和绿洲资本作为种子轮老股东持续加码；11月，获得数千万元天使+轮融资，领投方柏睿资本是由宁德时代联合创始人、副董事长李平出资创办的投资机构。2025年3月，千寻智能宣布完成5.28亿元Pre-A轮融资，阿美风险投资旗下 Prosperity7 Ventures（P7）领投，招商局创投、广发信德、靖亚资本、东方富海、华控基金等多家资本深度参与；7月，完成近6亿元PreA+轮融资，由京东领投，中国互联网投资基金（简称“中网投”）、浙江省科创母基金、华泰紫金、复星锐正等知名机构跟投。
2024年1月，千寻智能（杭州）科技有限公司正式注册成立。8月，获得近2亿元种子轮+天使轮融资，由弘晖基金领投，达晨创投、千乘资本跟投，同时，顺为资本和绿洲资本作为种子轮老股东持续加码；11月，获得数千万元天使+轮融资，领投方柏睿资本是由宁德时代联合创始人、副董事长李平出资创办的投资机构。2025年3月，千寻智能宣布完成5.28亿元Pre-A轮融资，阿美风险投资旗下 Prosperity7 Ventures（P7）领投，招商局创投、广发信德、靖亚资本、东方富海、华控基金等多家资本深度参与；7月，完成近6亿元PreA+轮融资，由京东领投，中国互联网投资基金…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
isaac-sim/IsaacLab v3.0.0-beta2.patch1e9ab2f866854b173
声明：## What's Changed * Bump h5py version to >=3.16.0 by @peterd-NV in https://github.com/isaac-sim/IsaacLab/pull/6266 * Updates Isaac Sim to version 6.0.1 by @kellyguo11 in https://github.com/isaac-sim/IsaacLab/pull/6277 * [Fix] Cherry-pick Isaac Sim 6.0 streaming crash fix by @hujc7 in https://github.com/isaac-sim/IsaacLab/pull/6295 **Full Changelog**: https://github.com/isaac-sim/IsaacLab/compare/v3.0.0-beta2...v3.0.0-beta2.patch1
## What's Changed * Bump h5py version to >=3.16.0 by @peterd-NV in https://github.com/isaac-sim/IsaacLab/pull/6266 * Updates Isaac Sim to version 6.0.1 by @kellyguo11 in https://github.com/isaac-sim/IsaacLab/pull/6277 * [Fix] Cherry-pick Isaac Sim 6.0 streaming crash fix by @hujc…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
How Toshio Fukuda Helped Shape IEEE and Global Roboticsec5a3339a514963b
声明：How Toshio Fukuda Helped Shape IEEE and Global Robotics
How Toshio Fukuda Helped Shape IEEE and Global Robotics
用世界模型给VLA当教练，原力灵机发布DW0.5，把RL搬进虚拟世界ecb4706cb834ee2f
声明：用世界模型给VLA当教练，原力灵机发布DW0.5，把RL搬进虚拟世界
用世界模型给VLA当教练，原力灵机发布DW0.5，把RL搬进虚拟世界
AeroAct: Action-Centered World-Action Models for Language-Conditioned Quadrotor Flightee35d9e64b9178f5
声明：To obtain aligned visual, state, language, and dynamically feasible action data, we build a DiffAero-based pipeline with complementary Isaac Lab and 3D Gaussian splatting renderers.
To obtain aligned visual, state, language, and dynamically feasible action data, we build a DiffAero-based pipeline with complementary Isaac Lab and 3D Gaussian splatting renderers.
A Steerable Model with Emergent Capabilitiesee7bf4a6aed30a53
声明：In our experiments, we see π0.7 exhibiting the first signs of compositional generalization, recombining skills from various tasks to solve new problems, like using new kitchen appliances and even enabling a new robot to fold laundry for which there is no laundry folding data.
In our experiments, we see π0.7 exhibiting the first signs of compositional generalization, recombining skills from various tasks to solve new problems, like using new kitchen appliances and even enabling a new robot to fold laundry for which there is no laundry folding data.
How Toshio Fukuda Helped Shape IEEE and Global Roboticsefe7fd3f26e6ab72
声明：Born in Japan, Fukuda has been recognized by the country for his contributions to science with two of its highest awards: the Medal of Honor with a purple ribbon in 2015 and the Order of the Sacred Treasure in 2022.
Born in Japan, Fukuda has been recognized by the country for his contributions to science with two of its highest awards: the Medal of Honor with a purple ribbon in 2015 and the Order of the Sacred Treasure in 2022.
用世界模型给VLA当教练，原力灵机发布DW0.5，把RL搬进虚拟世界f1f5d329219ffea4
声明：纸上谈兵终觉浅。可以看到，在打气球、晾衣服、叠纸盒等高难度复杂任务中，接入DFOL 2.0的模型相比单纯SFT（监督微调）基线，关键步骤成功率骤然提升。
纸上谈兵终觉浅。可以看到，在打气球、晾衣服、叠纸盒等高难度复杂任务中，接入DFOL 2.0的模型相比单纯SFT（监督微调）基线，关键步骤成功率骤然提升。
手握英伟达宁德时代相同剧本，具身智能的第一个「基建商」出现了f2b1a0e6aa1c1352
声明：具身智能热了三年，融资一轮比一轮大，但整体进展远不如预期：2026年过半，我们仍然在等待物理AI的“ChatGPT时刻”。
具身智能热了三年，融资一轮比一轮大，但整体进展远不如预期：2026年过半，我们仍然在等待物理AI的“ChatGPT时刻”。
DriftWorld: Fast World Modeling through Driftingf3108b4e53853125
声明：DriftWorld: Fast World Modeling through Drifting
DriftWorld: Fast World Modeling through Drifting
RoboTTT: Context Scaling for Robot Policiesf3896b31c8eaa6b8
声明：On challenging real-robot manipulation tasks, RoboTTT improves overall performance by 87% over the single-step context baseline and fully completes a five-minute, ten-stage assembly task, which no baseline ever does.
On challenging real-robot manipulation tasks, RoboTTT improves overall performance by 87% over the single-step context baseline and fully completes a five-minute, ten-stage assembly task, which no baseline ever does.
CosFly-VLA: A Spatially Aware Vision-Language-Action Model for UAV Trackingf5573adb47972297
声明：Spatially Grounded Continued Pretraining (CPT) on a 500k mixed pool injects UAV-view depth, distance, and 3-D spatial reasoning.
Spatially Grounded Continued Pretraining (CPT) on a 500k mixed pool injects UAV-view depth, distance, and 3-D spatial reasoning.
Building the general-purpose robotic brainf5f036975c714480
声明：Our foundation model, the Skild Brain, which works on various quadrupeds, humanoids, table-top arms, mobile manipulators and more, follows a hierarchical architecture: (1) a low-frequency high-level manipulation and navigation action policy which provides inputs to a (2) high-frequency low-level action policy.
Our foundation model, the Skild Brain, which works on various quadrupeds, humanoids, table-top arms, mobile manipulators and more, follows a hierarchical architecture: (1) a low-frequency high-level manipulation and navigation action policy which provides inputs to a (2) high-fre…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
Inside X Square Robot’s Bold Plan for Real Home Robotsf6d1516467e8c462
声明：X Square Robot’s valuation has climbed above 20 billion yuan (about US $2.9 billion), suggesting that investors increasingly view data infrastructure, foundation models, and scalable training systems as long-term differentiators in embodied AI.
X Square Robot’s valuation has climbed above 20 billion yuan (about US $2.9 billion), suggesting that investors increasingly view data infrastructure, foundation models, and scalable training systems as long-term differentiators in embodied AI.
huggingface/lerobot Release v0.5.0f7f5a4454c11c621
声明：## What's Changed * chore(dependencies): Bump lerobot to 0.4.4 by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/2840 * feat(ci): trigger manually documentation release version by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/2841 * fix(robots): update gripper configuration and calibration settings for OMX by @Woojin-Crive in https://github.com/huggingface/lerobot/pull/2815 * feat(envs): Add task ids to libero env cfg by @reeceomahoney in https://github.com/huggingface/lerobot/pull/2842 * feat(motors): add damiao motors & can bus by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/2788 * Add robot images to docs by @pkooij in https://github.com/
## What's Changed * chore(dependencies): Bump lerobot to 0.4.4 by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/2840 * feat(ci): trigger manually documentation release version by @imstevenpmwork in https://github.com/huggingface/lerobot/pull/2841 * fix(robots): u…
公开页面仅展示必要证据摘录；请访问原文核验完整上下文。
Building the general-purpose robotic brainfa7f0f2e3d599f86
声明：Thus, a lot of researchers and competitors side-step the problem: start with an existing vision-and-language model (VLM) and sprinkle in less than 1% of real-world robot data to build a “robotics foundation model”.
Thus, a lot of researchers and competitors side-step the problem: start with an existing vision-and-language model (VLM) and sprinkle in less than 1% of real-world robot data to build a “robotics foundation model”.
How Cheap Ground Robots Are Rewriting Frontline Warfare in Ukraineff2764b4d573573f
声明：How Cheap Ground Robots Are Rewriting Frontline Warfare in Ukraine
How Cheap Ground Robots Are Rewriting Frontline Warfare in Ukraine