cs.RO（2026-05-12）

📊 共 30 篇论文 | 🔗 7 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (23 🔗6) 支柱九：具身大模型 (Embodied Foundation Models) (4) 支柱二：RL算法与架构 (RL & Architecture) (3 🔗1)

🔬 支柱一：机器人控制 (Robot Control) (23 篇)

#题目一句话要点标签🔗⭐
1 Real-Time Whole-Body Teleoperation of a Humanoid Robot Using IMU-Based Motion Capture with Sim2Sim and Sim2Real Validation 提出基于IMU的全身动作捕捉系统，实现人形机器人实时遥操作 humanoid humanoid robot sim-to-real
2 World Action Models: The Next Frontier in Embodied AI 提出世界行动模型（WAMs）框架，统一预测状态建模与动作生成，提升具身智能 teleoperation policy learning world model
3 Morphologically Equivariant Flow Matching for Bimanual Mobile Manipulation 提出形态等变流匹配方法以解决双手移动操控问题 manipulation mobile manipulation bi-manual
4 RIO: Flexible Real-Time Robot I/O for Cross-Embodiment Robot Learning RIO：用于跨具身机器人学习的灵活实时机器人I/O框架 humanoid bi-manual teleoperation
5 DexTwist: Dexterous Hand Retargeting for Twist Motion via Mixed Reality-based Teleoperation DexTwist：基于混合现实遥操作的灵巧手旋转运动重定向 manipulation dexterous hand teleoperation
6 See What Matters: Differentiable Grid Sample Pruning for Generalizable Vision-Language-Action Model 提出可微网格采样器GridS，用于通用视觉-语言-动作模型中的几何感知型token剪枝。 manipulation vision-language-action VLA ✅
7 DreamAvoid: Critical-Phase Test-Time Dreaming to Avoid Failures in VLA Policies DreamAvoid：关键阶段测试时Dreaming，避免VLA策略失效 manipulation vision-language-action VLA ✅
8 Learning Action Manifold with Multi-view Latent Priors for Robotic Manipulation 针对机器人操作，提出基于多视角隐变量先验的动作流形学习方法 manipulation vision-language-action VLA ✅
9 Mapping Embodied Affective Touch Strategies on a Humanoid Robot 研究具身情感触摸策略在人型机器人上的映射，揭示身体区域依赖性和约束影响。 humanoid humanoid robot
10 NavOL: Navigation Policy with Online Imitation Learning NavOL：基于在线模仿学习的导航策略，解决分布偏移和奖励设计难题。 domain randomization reinforcement learning imitation learning
11 Offline Policy Evaluation for Manipulation Policies via Discounted Liveness Formulation 提出基于Liveness的折扣策略评估方法，解决机械臂操作策略的离线评估偏差问题 manipulation diffusion policy vision-language-action
12 From Reaction to Anticipation: Proactive Failure Recovery through Agentic Task Graph for Robotic Manipulation AgentChord：通过Agentic任务图实现机器人操作的主动故障恢复 manipulation bi-manual bimanual manipulation
13 SafeManip: A Property-Driven Benchmark for Temporal Safety Evaluation in Robotic Manipulation 提出SafeManip以解决机器人操作中的时间安全评估问题 manipulation vision-language-action
14 TMRL: Diffusion Timestep-Modulated Pretraining Enables Exploration for Efficient Policy Finetuning TMRL：扩散时间步调制预训练，提升探索能力，实现高效策略微调 manipulation reinforcement learning VLA ✅
15 X-Imitator: Spatial-Aware Imitation Learning via Bidirectional Action-Pose Interaction X-Imitator：通过双向动作-姿态交互实现空间感知模仿学习 manipulation imitation learning
16 Cooperative Robotics Reinforced by Collective Perception for Traffic Moderation 提出基于集体感知增强的合作机器人交通疏导系统，解决非视距路口安全问题 humanoid humanoid robot penetration
17 From Imagined Futures to Executable Actions: Mixture of Latent Actions for Robot Manipulation 提出MoLA：利用混合潜在动作弥合视频预测与机器人操作之间的鸿沟 manipulation
18 A Proprioceptive-Only Benchmark for Quadruped State Estimation: ATE, RPE, and Runtime Trade-offs Between Filters and Smoothers 四足机器人本体感知状态估计基准测试：滤波器与平滑器在ATE、RPE和运行时间上的权衡 quadruped ✅
19 Sampling-Based Follow-the-Leader Motion Planning for Manipulator-Mounted Continuum Robots 提出基于采样的随动控制运动规划算法，用于机械臂搭载的连续体机器人。 motion planning ✅
20 Rainbow Deep Q-Learning with Kinematics-Aware Design for Cooperative Delta and 3-RRS Parallel Robot Insertion 提出一种结合运动学感知的Rainbow DQN，用于Delta和3-RRS并联机器人协同装配 manipulation reinforcement learning deep reinforcement learning
21 Coordinated Diffusion: Generating Multi-Agent Behavior Without Multi-Agent Demonstrations CoDi：利用单智能体数据生成多智能体协调行为，无需多智能体演示 manipulation imitation learning diffusion policy
22 Learning What Matters: Adaptive Information-Theoretic Objectives for Robot Exploration 提出QOED自适应信息目标，提升机器人探索中模型参数学习的效率与准确性 manipulation reinforcement learning
23 Introducing Environmental Constraints to Grasping Strategies for Paper-Like Flexible Materials Using a Soft Gripper 针对纸状柔性材料，提出利用环境约束的软抓取策略 manipulation

🔬 支柱九：具身大模型 (Embodied Foundation Models) (4 篇)

#题目一句话要点标签🔗⭐
24 Premover: Fast Vision-Language-Action Control by Acting Before Instructions Are Complete Premover：通过指令未完成前的预执行加速视觉-语言-动作控制 vision-language-action VLA
25 Overcoming Dynamics-Blindness: Training-Free Pace-and-Path Correction for VLA Models 提出免训练的Pace-and-Path Correction方法，解决VLA模型在动态场景下的适应性问题 vision-language-action VLA
26 GuidedVLA: Specifying Task-Relevant Factors via Plug-and-Play Action Attention Specialization GuidedVLA：通过插件式动作注意力特化引导视觉-语言-动作模型 vision-language-action VLA
27 Kairos: A Scalable Serving System for Physical AI Kairos：一种可扩展的物理人工智能服务系统，优化生成-执行循环。 foundation model

🔬 支柱二：RL算法与架构 (RL & Architecture) (3 篇)

#题目一句话要点标签🔗⭐
28 EvoNav: Evolutionary Reward Function Design for Robot Navigation with Large Language Models EvoNav：利用大语言模型进化式设计机器人导航奖励函数 reinforcement learning reward design large language model
29 SI-Diff: A Framework for Learning Search and High-Precision Insertion with a Force-Domain Diffusion Policy SI-Diff：力域扩散策略学习搜索与高精度插入，解决机器人装配中的位姿不确定性问题。 diffusion policy zero-shot transfer
30 Closing the Motion Execution Gap: From Semantic Motion Task Constraints to Kinematic Control 提出Giskard框架，利用运动状态图和lMPC解决语义约束到运动控制的难题 world model world models cross-embodiment ✅

⬅️ 返回 cs.RO 首页 · 🏠 返回主页