Multi-UAV State Estimation专业
融合倾斜测量的线性推力约束卡尔曼滤波提升多无人机敏捷飞行状态估计
这是一篇预印本。作者针对视觉多无人机飞行中仅用位置测量导致高阶状态估计延迟的问题，提出融合视觉检测器提供的倾斜信息，并设计了一种线性推力约束卡尔曼滤波器。在三个数据集上，姿态感知估计将速度和加速度误差平均降低40%和57%，闭环仿真中支持超过2g的横向机动跟踪。
研究
论文
机器人与具身智能领域的最新论文，以及每篇对产业的意义。
暂无中文版，显示英文原文。
VLA进阶
ChunkTrust：利用动作专家证据自适应调整机器人策略的执行时域
机器人AI模型每次预测一段未来动作，执行多少步之后再重新观察环境，过去大多由工程师手动设为固定值。清华大学、北京智源人工智能研究院（BAAI）等机构发布的这篇预印本提出一个即插即用模块，在运行中自动决定这一步数，无需重新训练，就提升了Physical Intelligence和英伟达现有模型在仿真和真实双臂机器人上的成功率。
Autonomous Exploration专业
GPU加速路径依赖边际信息增益方法，无人机探索效率提升30%
本文为预印本。作者提出一种GPU加速的路径依赖边际信息增益计算方法，用深度缓冲区表示祖先观测，在GPU上按深度顺序评估规划树，避免重复计算重叠区域。相比精确体素哈希映射方法，该法误差在5%至10%以内，桌面GPU最高加速118倍，在仿真和真机测试中分别使95%覆盖率时间缩短最多3.8分钟和30%。
Model-based RL专业
PL-MPC 在模型预测控制中同时改进 critic 监督、终端估计与策略蒸馏，真机 unseen 尺寸成功率超 TD-M(PC)2
PL-MPC 是 arXiv 预印本，未注明经同行评审。该方法在 TD-M(PC)2 上同步修改 critic 监督、MPC 终端值估计和 planner 到策略的转移，不改世界模型与 MPPI。在 HumanoidBench 的 balance-hard 上 TAR 从 98±18 升至 387±255，hurdle 从 199±13 升至 466±200；在 KUKA IIWA14 真机 wrench–nut 对齐中，训练尺寸和两个未见尺寸的观测成功率均高于 TD-M(PC)2。
Trajectory Correction进阶
MaDE：冻结后处理算子将神经轨迹预测强制到学习动力学流形，显著降低残差但位移误差上升
论文提出Markovian Dynamics Enforcer（MaDE），一个无需真值控制、时间不变的后处理算子，可为任意神经轨迹预测器附加动力学一致性与不等式约束修正。在inD真实车辆轨迹上，MaDE将一步运动学自行车模型残差从原始预测器的0.1703至0.1714降至0.0071至0.0072，但平均位移误差增大1.57至1.83倍。该工作为arXiv预印本，未经同行评审。
Terrain traversability专业
经验驱动的持续学习预测四足机器人地形可通过性，验证门控降低遗忘23.1%
这篇预印本提出一个持续学习管线，利用冻结的DINOv3视觉特征和证据回归器从接触前图像预测五种足端交互指标，并通过有界重放和验证门控持续更新。在Unitree Go2真机顺序经过三种新地形时，带历史验证门控的方法比仅重放方法将最终锚点负对数似然退化降低23.1%，同时新地形适应相近。该预印本尚未经同行评审。
Real-time VLA专业
两阶段非均匀Flow去噪把VLA推理从61.6毫秒压到22.0毫秒，真机叠衣验证实时执行方法
这是一篇arXiv预印本，尚未经过同行评审。作者以π0.5为基线，在双臂真机上测量了从相机、本体感知到命令响应的端到端延迟，并提出两阶段非均匀Flow Matching去噪，把推理步数从10步减到2步，模型推理时间从61.557毫秒降到21.956毫秒。在30次叠衣真机试验中，训练型方法Legato和免训练方法Temporal Smoothing表现最好；把两阶段去噪与这两种方法结合，推理成本大幅下降，但任务成功率有小幅回落。
Text-to-3D Policy进阶
T3DP：细粒度语言-行为对齐提升文本到3D策略的未见过规格泛化
本文提出T3DP框架，通过在指令token和局部行为表示之间建立双向细粒度对齐，让3D扩散策略能执行训练中未出现的精细行为规格。在Meta-World、ManiSkill和RoboTwin上，T3DP较全局语言-行为对齐平均成功率提升11.0至14.2个百分点；在真实机器人上平均成功率从47.5%升至65.0%。该工作为预印本，尚未经过同行评审。
Safe RL-MPC专业
RL引导PAC-NMPC：固定翼无人机视觉导航实现概率安全与远见
本文提出将强化学习训练的actor-critic和传感器预测模型嵌入PAC-NMPC框架，为未知环境中基于感知的导航提供有限时间概率安全保证和长期性能。在固定翼无人机上，该方法在仿真中取得90%成功率，在硬件试验中取得80%成功率，均优于纯RL和传统规划基线。本文为预印本，未经同行评审。
Embodied policy专业
EWAM统一具身模型实现语义、视觉预测、动作的内部递进分工
EWAM是一种以动作流为核心的一体化具身模型，通过非对称联合注意力让动作token在每层读取语义、当前视觉、预测未来和动作信息。在RoboTwin clean-to-random测试中达到72.1%，LIBERO平均98.8%，并在多个真实机器人上超过VLA、WAM和混合基线。该预印本显示统一训练可诱导从语义理解、视觉预测到动作生成的内部有序计算，且人类自我中心数据可提升跨具身迁移和真实场景鲁棒性。
Social Navigation进阶
Social-WM：用可执行动作差异做安全社交导航的隐世界模型
Social-WM 是一个使用 RGB 视频离线训练的隐世界模型规划框架，区分名义动作与实际可执行动作。在 Social-HM3D 上，它取得 63.77% 成功率，行人碰撞率降至 21.67%，比 NavThinker 相对降低 44.6%；零样本迁移到 Social-MP3D 仍保持 54.15% 成功率和 23.34% 碰撞率。该工作为未同行评审的预印本。
Manipulation进阶
RoboCoach 让世界模型主动当教练：想象失败后定向补数据，真机成功率升至 75.0% 和 83.8%
RoboCoach 是一个预印本框架，用世界模型中的想象执行来诊断长程机器人任务中最先失败的子任务，并据此请求对应技能专家的演示数据。在 Franka 和 AgileX 真机上，仅追加 150 条子任务演示，成功率分别从 13.3% 升至 75.0%、从 40.0% 升至 83.8%，显著超过统一采集加共享适配器的基线。该工作表明世界模型可以成为主动教练，把稀缺的真实数据导向可复用的技能模块。
Navigation专业
Comparing Utility of Inertial, Occupancy, Semantic, and Intent Information in Human Motion Prediction During Daily Tasks
Accurate human motion prediction is crucial for robotic systems operating around people, particularly in complex indoor spaces. In this study, we assess the relative importance of different sources of information in indoor motion prediction with a human motion diffusion model.
Perception专业
ProAct-VLM: Pre-Failure Vision-Language Task Replanning with Continuous Perception Feedback
Long-horizon robotic tasks are vulnerable to unexpected environmental changes that can render planned actions ineffective or unsafe. To address this, robots must detect such changes as they occur, interpret their impact, and adjust their actions accordingly.
Manipulation专业
MVG-WAM: Multiple View Geometry-Aware World-Action Modeling for Robotic Manipulation
World-Action Models (WAMs) couple visual dynamics with action prediction, bringing the rich priors of pretrained video models to robotic manipulation. However, their multi-view interfaces typically tile images or concatenate tokens, leaving the geometric relationships among synchronized cameras implicit.
VLA专业
MotorMind: Scaffolding General Vision Language Models for Zero-Shot Robot Manipulation
Vision-language-action (VLA) models have advanced robotic manipulation, but their zero-shot generalization in new tasks and environments remains limited, and their reliance on specialized training keeps them from benefiting directly from rapidly advancing general-purpose vision-language models (VLMs). In parallel, recent agentic robotic systems leverage VLMs for high-level reasoning or coding agents for robot control, but often depend on extensive external models and tools, introducing additional complexity and cost.
Manipulation专业
BlenDAgger: Blended Shared Control for Interactive Imitation Learning
Robot policies are frequently trained from human corrections, yet teleoperating a robot to provide corrections is burdensome, and human demonstrators are not always optimal. We propose Blended DAgger (BlenDAgger), an approach for collecting data to train imitation learning policies by using shared control to blend the policy's and demonstrator's actions during interventions.
Humanoid专业
EgoAlign: Bridging the Human-Humanoid Gap for Long-Range Loco-Manipulation
Egocentric human demonstrations offer an accessible source of task experience, but differences in body scale and controller response, together with missing robot states, limit their value as humanoid training supervision. We present EgoAlign, a data-construction framework that converts these demonstrations into action and state supervision compatible with a general-purpose, continuous whole-body controller, without collecting physical-robot demonstrations.
Learning专业
Rethinking Representations for World-Action Modeling
World-action models jointly learn robot policies and predict future observations, making the representation space an interface between control and prediction. We study the design of this space through controlled comparisons, finding that neither reconstruction fidelity nor pre-trained perceptual features alone ensure effective policy learning.
Navigation专业
Geometry-Preserving Human-to-Robot Upper-Body Motion Retargeting from Monocular Video
Monocular RGB video provides an accessible source of human demonstrations for upper-body robot motion, yet video-driven human-to-robot transfer remains challenging because body and hand motion are recovered at different spatial scales, human and robot kinematics differ substantially, and fine distal motion is difficult to preserve across embodiments. We present a geometry-preserving motion-retargeting framework that integrates unified body--hand reconstruction with morphology-independent geometric transfer.
主题页
论文来自 arXiv 机器人领域。尚未生成摘要的论文，显示摘要开头部分。