cs.RO（2025-05-16）

📊 共 25 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (16 🔗4) 支柱九：具身大模型 (Embodied Foundation Models) (4) 支柱二：RL算法与架构 (RL & Architecture) (2) 支柱三：空间感知与语义 (Perception & Semantics) (2 🔗1) 支柱八：物理动画 (Physics-based Animation) (1)

🔬 支柱一：机器人控制 (Robot Control) (16 篇)

#题目一句话要点标签🔗⭐
1 Parkour in the Wild: Learning a General and Extensible Agile Locomotion Policy Using Multi-expert Distillation and RL Fine-tuning 提出基于多专家蒸馏与强化学习微调的通用敏捷足式机器人运动策略 legged robot locomotion locomotion policy
2 Bracing for Impact: Robust Humanoid Push Recovery and Locomotion with Reduced Order Models 提出基于降阶模型的统一框架，实现人oids在行走过程中的鲁棒抗扰和平衡恢复。 humanoid humanoid robot locomotion
3 GRoQ-LoCO: Generalist and Robot-agnostic Quadruped Locomotion Control using Offline Datasets GRoQ-LoCO：基于离线数据的通用且机器人无关的四足机器人运动控制 quadruped legged locomotion locomotion
4 Learning Multimodal AI Algorithms for Amplifying Limited User Input into High-dimensional Control Space 提出ARAS框架，利用多模态AI和强化学习，将有限用户输入转化为高维控制信号，辅助瘫痪患者。 manipulation dexterous manipulation sim-to-real
5 DexGarmentLab: Dexterous Garment Manipulation Environment with Generalizable Policy 提出DexGarmentLab环境和HALO策略，解决服装灵巧操作中泛化性问题。 manipulation bi-manual sim-to-real ✅
6 Zero-Shot Visual Generalization in Robot Manipulation 提出基于解耦表示学习和联想记忆的机器人操作零样本视觉泛化方法 manipulation domain randomization reinforcement learning
7 Grounded Task Axes: Zero-Shot Semantic Skill Generalization via Task-Axis Controllers and Visual Foundation Models 提出基于任务轴控制和视觉基础模型的零样本语义技能泛化方法 manipulation foundation model zero-shot transfer
8 SHIELD: Safety on Humanoids via CBFs In Expectation on Learned Dynamics SHIELD：基于学习动力学期望CBF的人形机器人安全保障 humanoid locomotion Unitree
9 ReWiND: Language-Guided Rewards Teach Robot Policies without New Demonstrations ReWiND：无需新示教，通过语言引导奖励学习机器人策略 manipulation bi-manual reinforcement learning ✅
10 X2C: A Dataset Featuring Nuanced Facial Expressions for Realistic Humanoid Imitation X2C：用于逼真人形模仿的细微面部表情数据集与模仿框架 humanoid humanoid robot human-to-humanoid ✅
11 Unleashing Humanoid Reaching Potential via Real-world-Ready Skill Space 提出R2S2技能空间，实现人型机器人大范围、零样本的真实环境抓取 humanoid whole-body control sim2real
12 Self-supervised perception for tactile skin covered dexterous hands 提出Sparsh-skin，用于灵巧手磁性皮肤触觉感知的自监督预训练编码器。 dexterous hand policy learning distillation
13 Exploiting Radiance Fields for Grasp Generation on Novel Synthetic Views 利用辐射场生成新视角下的抓取姿态，提升机器人操作性能 manipulation gaussian splatting splatting
14 Object-Centric Representations Improve Policy Generalization in Robot Manipulation 提出基于对象中心表示的机器人操作策略，提升策略泛化能力 manipulation
15 Estimating Deformable-Rigid Contact Interactions for a Deformable Tool via Learning and Model-Based Optimization 提出一种混合学习与模型优化的方法，用于估计可变形工具与刚性物体的接触交互 manipulation dexterous manipulation ✅
16 Decoupling Collision Avoidance in and for Optimal Control using Least-Squares Support Vector Machines 利用最小二乘支持向量机解耦最优控制中的避碰问题 motion planning

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

#题目一句话要点标签🔗⭐
17 Unveiling the Potential of Vision-Language-Action Models with Open-Ended Multimodal Instructions 提出OE-VLA模型，扩展VLA模型至开放式多模态指令机器人控制 vision-language-action VLA foundation model
18 Search-TTA: A Multimodal Test-Time Adaptation Framework for Visual Search in the Wild 提出Search-TTA框架以解决户外视觉搜索中的不确定性问题 multimodal
19 Real-Time Verification of Embodied Reasoning for Generative Skill Acquisition 提出VERGSA框架，通过实时验证提升具身智能体生成式技能学习效率。 generalist agent
20 REI-Bench: Can Embodied Agents Understand Vague Human Instructions in Task Planning? 提出REI-Bench以解决机器人任务规划中的模糊人类指令问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (2 篇)

#题目一句话要点标签🔗⭐
21 Conditioning Matters: Training Diffusion Policies is Faster Than You Think Cocos：通过条件依赖的扩散策略训练加速视觉-语言-动作模型的学习。 diffusion policy flow matching vision-language-action
22 Counterfactual Behavior Cloning: Offline Imitation Learning from Imperfect Human Demonstrations 提出Counter-BC，从不完美的人类演示中进行离线模仿学习。 imitation learning behavior cloning

🔬 支柱三：空间感知与语义 (Perception & Semantics) (2 篇)

#题目一句话要点标签🔗⭐
23 Planar Velocity Estimation for Fast-Moving Mobile Robots Using Event-Based Optical Flow 提出基于事件相机的平面速度估计方法，提升快速移动机器人速度估计精度。 VIO optical flow
24 GrowSplat: Constructing Temporal Digital Twins of Plants with Gaussian Splats GrowSplat：利用高斯溅射构建植物生长过程的四维数字孪生模型 3D gaussian splatting gaussian splatting splatting ✅

🔬 支柱八：物理动画 (Physics-based Animation) (1 篇)

#题目一句话要点标签🔗⭐
25 Adaptive Ergodic Search with Energy-Aware Scheduling for Persistent Multi-Robot Missions 提出mEclares框架，通过自适应遍历搜索和能量感知调度解决多机器人持久任务问题。 spatiotemporal

⬅️ 返回 cs.RO 首页 · 🏠 返回主页