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。
Translating language into humanoid motion requires coordinating body parts, determining motion duration, and producing references suitable for robot execution. We introduce PartiGen, a part-aware latent diffusion framework that incorporates anatomical structure into a shared whole-body motion representation.
End-to-end planning in complex scenes must model diverse driving intentions and select the most appropriate trajectory using scene context and future risks. However, multi-query regression may collapse to one mode, while anchor-based and generative planners share a scaling bottleneck: small candidate sets limit maneuver coverage or discovery of high-quality trajectories, whereas large sets increase generation and scoring costs.
Robot policies often convert sensing and control into fixed-rate observation-action sequences, leaving physical time implicit in sequence position. This assumption breaks when observation rates or delays vary, or when inference latency changes.
A user who brings a generalist manipulation policy into their home or office wants it to pick up their objects from their shelves, counters, and tables. Yet even a strong generalist like $π_{0.5}$ achieves 0% pickup success in our captured deployment scenes without adaptation, and collecting real demonstrations in every user's environment is prohibitively expensive.
Nearly every autonomous or semi-autonomous flipper controller assumes access to an elevation map of the terrain under and in front of the robot. The terrain beneath the chassis is occluded in the sensors, so the map is remembered through odometry integration, whose drift corrupts it.
Autonomous landing of unmanned aerial vehicle (UAV) on unmanned surface vehicle (USV) is difficult in high sea states. Existing hierarchical landing methods often infer future USV motion only implicitly from short observation histories, which limits their ability to adapt to severe and rapidly changing sea conditions.
Humanoid tennis requires reactive whole-body coordination under rapidly changing ball dynamics, yet collecting task-specific motion-capture demonstrations that cover diverse approach states, contact locations, and timing conditions is costly. Inspired by how humans acquire new skills by combining reusable motor experience with task-specific demonstrations, we present ATHLETE, a framework for learning reactive humanoid tennis from generic locomotion data and ordinary tennis videos.
Generating edge case scenarios is crucial for training, testing and safety assuring autonomous systems. This identifies hazardous failures to increase system reliability and robustness, ensures the system can still function safely on the boundaries of its operational design domain and learns the expected system behaviour for these edge cases.
Scaling generalist robotic policies is limited by the high cost of collecting embodied interaction data. Large-scale egocentric human videos offer a rich alternative, but they lack robot action labels, and existing action annotation methods are often noisy under real-world visual complexity and diverse manipulation dynamics.
World Action Models (WAMs) inherit powerful priors from pretrained video generation models, but their variational autoencoder (VAE) representations are optimized for visual reconstruction rather than manipulation. External visual representations can improve manipulation, but existing methods couple representation choices with different integration mechanisms, obscuring their individual effects.
本文提出 SafeStreamingFlow，一种从演示中学习的目标条件流式规划器。它把流采样时间与物理执行时间对齐，仅对执行步施加高阶控制屏障函数约束，将优化求解次数从 O(HK) 降至 O(H)。在 Maze2D、F1TENTH、Hopper 和 Gazebo 仓库等仿真基准上，该方法在保持目标到达竞争力的同时，把执行期违规率降至接近零，并把规划延迟降低一个数量级。本工作为预印本，尚未经过同行评审。
这篇预印本提出将 empowerment 直接作为采样式模型预测控制（MPC）的目标，用同一个动作分布完成探测与控制。在经典控制任务上，结合 empowerment 与稀疏任务代价后，成功率最高从单独 empowerment 的 59% 提升到 100%，而仅用稀疏代价的基线全部失败。该方法只需一阶导数，为 MPC 引入内在动机提供了实用桥梁。
本预印本提出 NAVA-WAM，一种世界动作模型，通过未来视频流匹配直接在无动作视频上预训练动作策略，避免了中间表示或潜在动作接口。在 LIBERO、RoboTwin 2.0 以及 Franka FR3 真机实验中，该方法分别达到 99.0%、83.5%、88.5%、73.6% 和 93.3% 的成功率，并在低动作标签预算下显著优于基线。该结果展示了从规模化无动作视频中直接学习动作先验的潜力，为降低对带动作标签机器人数据的依赖提供新路径。
SimpleTouch 是一篇 arXiv 预印本。它给 π0.5 增加一个触觉专家，读取冻结 T3 编码器的全部 token，只用任务演示和未来触觉预测监督训练，无需触觉策略预训练或单独对齐。在六项 UniVTAC 任务上平均成功率77.5%，比 FTP-1 高10.8个百分点；在四项真机任务上平均71.3%，比 FTP-1 高8.8个百分点。
本文报告一种预印本方法 DR-IPC，它结合 EKF 状态滤波与非线性扰动观测器，将扰动估计注入 NMPC 预测，联合优化局部避障轨迹和控制量。在 Gazebo 多目标导航中，DR-IPC 在有风扰和悬挂载荷时将任务完成率从 1/10 提高到 9/10；室内拱门实验中高度 RMSE 从 0.34 m 降至 0.01 m。该方法尚未经过同行评审，代码计划开源但尚未发布。
RoboBridge 是一个自进化具身智能体框架，把仿真到真实迁移看作可执行任务技能的持续适应。它把预训练 VLA 策略作为可调用动作工具，结合推理时引导和反馈驱动的技能增删改验证，在 LIBERO-PRO 上总体成功率 74.4%，超过最强基线 Harness VLA 的 72.1%；在三个真实机器人任务上，仿真技能直接迁移后经真实反馈进化，成功率分别达到 15/15、14/15、15/15。该论文为预印本，尚未经过同行评审。
这篇arXiv预印本尚未经过同行评审。作者提出UniIntervene++，一种在真实机器人在线强化学习中动态分配自主执行与多种辅助行为的干预智能体。它在五个真实操作任务上平均成功率达89.67%，比最强基线高6个百分点，人类介入率仅0.77%，比基线中最低的14.29%相对降低94.6%。
本文介绍预印本 SceneFactory-3D，一个 GPU 批量、物理接地的多智能体驾驶模拟器，能在匹配的 3D 反事实世界中仅改变路面摩擦或坡度，评估同一场景和控制器在不同道路条件下的闭环表现。实验覆盖 21 种摩擦与坡度条件、1024 个 12 车世界，结果显示摩擦从 1.0 降至 0.18 时，学习型策略安全通过率下降 6 至 90 个百分点，经典规划器下降 18 至 19 个百分点，且近碰撞事件更频繁。该框架为自动驾驶安全评估提供了可扩展的物理敏感性分析工具。
该预印本提出 OmniAct3D，一个将透视训练的基础模型3D检测器适配到全景等距柱状投影（ERP）图像的框架。在 Spheriverse 和 PanoMMOcc 基准上，OmniAct3D 分别达到 0.1887 mAP/0.1825 NDS 和 0.5897 mAP/0.4032 NDS，NDS 较先前最佳检测器提升 2.96 点，mAP 较未适配基础模型基线提升 24.87 点。论文尚未经同行评审，其核心贡献是通过几何适配与证据驱动的推理链保留基础模型先验并改善全景3D检测。
HexVIO将Basalt立体惯性里程计的视觉前端卸载到高通Hexagon DSP上，在三星Galaxy S23上实现了功耗降低67%或吞吐量提升86%。在40分钟测试中，以0.83W功耗维持30fps实时跟踪，相当于15Wh电池可运行约18小时。该研究为预印本，尚未经过同行评审。
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。