cs.RO（2025-09-23）

📊 共 45 篇论文 | 🔗 8 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (36 🔗6) 支柱二：RL算法与架构 (RL & Architecture) (5 🔗1) 支柱九：具身大模型 (Embodied Foundation Models) (2 🔗1) 支柱六：视频提取与匹配 (Video Extraction) (1) 支柱三：空间感知与语义 (Perception & Semantics) (1)

🔬 支柱一：机器人控制 (Robot Control) (36 篇)

#题目一句话要点标签🔗⭐
1 RoMoCo: Robotic Motion Control Toolbox for Reduced-Order Model-Based Locomotion on Bipedal and Humanoid Robots RoMoCo：用于双足和人形机器人基于降阶模型的运动控制工具箱 humanoid humanoid robot bipedal
2 Eva-VLA: Evaluating Vision-Language-Action Models' Robustness Under Real-World Physical Variations Eva-VLA：评估视觉-语言-动作模型在真实物理变化下的鲁棒性 manipulation scene understanding vision-language-action
3 Agentic Scene Policies: Unifying Space, Semantics, and Affordances for Robot Action 提出Agentic Scene Policies以解决复杂指令下的机器人动作问题 manipulation motion planning imitation learning
4 Reduced-Order Model-Guided Reinforcement Learning for Demonstration-Free Humanoid Locomotion 提出基于降阶模型引导的强化学习方法，实现无需演示的人形机器人运动控制 humanoid humanoid locomotion locomotion
5 MV-UMI: A Scalable Multi-View Interface for Cross-Embodiment Learning MV-UMI：用于跨具身学习的可扩展多视角交互界面 manipulation teleoperation imitation learning
6 Pure Vision Language Action (VLA) Models: A Comprehensive Survey VLA模型综述：将视觉语言模型从序列生成器转变为机器人控制的主动Agent manipulation vision-language-action VLA
7 World4RL: Diffusion World Models for Policy Refinement with Reinforcement Learning for Robotic Manipulation World4RL：利用扩散世界模型和强化学习改进机器人操作策略 manipulation sim-to-real reinforcement learning ✅
8 VGGT-DP: Generalizable Robot Control via Vision Foundation Models 提出VGGT-DP，利用视觉基础模型提升机器人操作技能的泛化性 manipulation imitation learning VGGT
9 EgoBridge: Domain Adaptation for Generalizable Imitation from Egocentric Human Data EgoBridge：利用领域自适应实现从第一视角人类数据中泛化模仿学习 manipulation bi-manual bimanual manipulation
10 ManipForce: Force-Guided Policy Learning with Frequency-Aware Representation for Contact-Rich Manipulation ManipForce：提出力引导的策略学习方法，用于接触式操作任务。 manipulation policy learning imitation learning
11 Chasing Stability: Humanoid Running via Control Lyapunov Function Guided Reinforcement Learning 提出基于控制Lyapunov函数引导的强化学习方法，实现人形机器人稳定奔跑 humanoid humanoid robot locomotion
12 Bi-VLA: Bilateral Control-Based Imitation Learning via Vision-Language Fusion for Action Generation 提出Bi-VLA，通过视觉-语言融合的模仿学习，解决机器人单模型多任务动作生成问题。 manipulation imitation learning VLA ✅
13 SINGER: An Onboard Generalist Vision-Language Navigation Policy for Drones SINGER：一种用于无人机的通用视觉-语言导航策略，仅使用机载传感器。 manipulation sim-to-real gaussian splatting
14 FUNCanon: Learning Pose-Aware Action Primitives via Functional Object Canonicalization for Generalizable Robotic Manipulation FUNCanon：通过功能对象规范化学习姿态感知动作原语，实现通用机器人操作 manipulation sim2real policy learning
15 Residual Off-Policy RL for Finetuning Behavior Cloning Policies 提出残差离线强化学习，微调行为克隆策略，实现高自由度机器人灵巧操作 humanoid humanoid robot manipulation
16 Imitation-Guided Bimanual Planning for Stable Manipulation under Changing External Forces 提出模仿引导的双臂规划框架，解决动态环境下稳定操作的抓取过渡问题。 manipulation bi-manual imitation learning
17 PIE: Perception and Interaction Enhanced End-to-End Motion Planning for Autonomous Driving PIE：面向自动驾驶，提出感知交互增强的端到端运动规划框架 motion planning Mamba scene understanding
18 ROPA: Synthetic Robot Pose Generation for RGB-D Bimanual Data Augmentation ROPA：用于RGB-D双臂操作数据增强的合成机器人姿态生成 manipulation bi-manual bimanual manipulation ✅
19 A Bimanual Gesture Interface for ROS-Based Mobile Manipulators Using TinyML and Sensor Fusion 提出基于TinyML和传感器融合的双手动势接口，用于ROS移动机械臂控制 manipulation bi-manual multimodal
20 3D Flow Diffusion Policy: Visuomotor Policy Learning via Generating Flow in 3D Space 提出3D FDP，通过生成3D空间中的Flow学习通用机器人操作策略。 manipulation policy learning diffusion policy
21 BiGraspFormer: End-to-End Bimanual Grasp Transformer BiGraspFormer：端到端双臂抓取Transformer网络，解决复杂物体操作中的协调问题。 manipulation bi-manual bimanual manipulation
22 Lang2Morph: Language-Driven Morphological Design of Robotic Hands Lang2Morph：提出一种基于语言驱动的机器人手部形态自动设计框架 manipulation dexterous hand human-object interaction
23 SPiDR: A Simple Approach for Zero-Shot Safety in Sim-to-Real Transfer SPiDR：一种基于悲观域随机化的简单零样本安全Sim-to-Real迁移方法 sim-to-real domain randomization reinforcement learning
24 Spatial Envelope MPC: High Performance Driving without a Reference 提出基于空间包络的MPC框架，无需参考轨迹实现高性能自动驾驶。 MPC model predictive control reinforcement learning
25 Growing with Your Embodied Agent: A Human-in-the-Loop Lifelong Code Generation Framework for Long-Horizon Manipulation Skills 提出人机协作的终身代码生成框架，提升长时程操作技能 manipulation large language model
26 Do You Need Proprioceptive States in Visuomotor Policies? 提出State-free策略，解决基于模仿学习的机器人操作中对本体感受状态的过度依赖问题。 manipulation whole-body manipulation imitation learning
27 Guaranteed Robust Nonlinear MPC via Disturbance Feedback 提出基于扰动反馈的鲁棒非线性MPC，保障机器人安全约束与稳定性 MPC model predictive control ✅
28 N2M: Bridging Navigation and Manipulation by Learning Pose Preference from Rollout 提出N2M以解决移动操作中导航与操作不一致问题 manipulation mobile manipulation
29 DexSkin: High-Coverage Conformable Robotic Skin for Learning Contact-Rich Manipulation 提出DexSkin以解决机器人触觉感知不足问题 manipulation reinforcement learning ✅
30 Distributionally Robust Safe Motion Planning with Contextual Information 提出一种基于上下文信息的分布鲁棒安全运动规划方法，用于解决复杂环境下的避障问题。 motion planning
31 MagiClaw: A Dual-Use, Vision-Based Soft Gripper for Bridging the Human Demonstration to Robotic Deployment Gap 提出MagiClaw以解决人类示范与机器人执行之间的领域差距问题 manipulation teleoperation policy learning
32 Spectral Signature Mapping from RGB Imagery for Terrain-Aware Navigation 提出RS-Net，利用RGB图像预测光谱特征，实现地形感知导航 quadruped MPC
33 Generalizable Domain Adaptation for Sim-and-Real Policy Co-Training 提出基于最优传输的Sim-to-Real策略协同训练框架，提升机器人操作泛化性 manipulation behavior cloning
34 Look as You Leap: Planning Simultaneous Motion and Perception for High-DOF Robots 提出基于神经代理模型的GPU并行感知评分引导概率路线图规划器，解决高自由度机器人运动感知协同规划问题。 motion planning
35 Crater Observing Bio-inspired Rolling Articulator (COBRA) COBRA：一种用于月球陨石坑探索的仿生滚动关节机器人 locomotion
36 SOE: Sample-Efficient Robot Policy Self-Improvement via On-Manifold Exploration SOE：基于流形探索的机器人策略自提升，提升采样效率与安全性 manipulation ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#题目一句话要点标签🔗⭐
37 LCMF: Lightweight Cross-Modality Mambaformer for Embodied Robotics VQA 提出轻量级跨模态Mambaformer（LCMF），用于具身机器人VQA任务。 Mamba SSM state space model
38 End-to-End Crop Row Navigation via LiDAR-Based Deep Reinforcement Learning 提出基于LiDAR和深度强化学习的端到端作物行导航方法 reinforcement learning deep reinforcement learning policy learning
39 Self-evolved Imitation Learning in Simulated World 提出自进化模仿学习框架SEIL，解决少样本模仿学习中专家数据匮乏问题。 imitation learning generalist agent ✅
40 Real-Time Reinforcement Learning for Dynamic Tasks with a Parallel Soft Robot 提出基于课程学习的最大扩散强化学习，实现软体机器人在动态任务中的实时单次部署控制。 reinforcement learning curriculum learning
41 Query-Centric Diffusion Policy for Generalizable Robotic Assembly 提出Query-centric Diffusion Policy，解决机器人装配中高层规划与底层控制的鸿沟 diffusion policy

🔬 支柱九：具身大模型 (Embodied Foundation Models) (2 篇)

#题目一句话要点标签🔗⭐
42 VLN-Zero: Rapid Exploration and Cache-Enabled Neurosymbolic Vision-Language Planning for Zero-Shot Transfer in Robot Navigation VLN-Zero：面向机器人导航零样本迁移的快速探索与缓存神经符号视觉语言规划 VLN zero-shot transfer ✅
43 A Multimodal Stochastic Planning Approach for Navigation and Multi-Robot Coordination 提出一种多模态随机规划方法，用于导航和多机器人协调。 multimodal

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#题目一句话要点标签🔗⭐
44 OmniVLA: An Omni-Modal Vision-Language-Action Model for Robot Navigation OmniVLA：用于机器人导航的通用模态视觉-语言-动作模型 egocentric vision-language-action VLA

🔬 支柱三：空间感知与语义 (Perception & Semantics) (1 篇)

#题目一句话要点标签🔗⭐
45 Terra: Hierarchical Terrain-Aware 3D Scene Graph for Task-Agnostic Outdoor Mapping Terra：面向任务无关户外建图的分层地形感知3D场景图 traversability

⬅️ 返回 cs.RO 首页 · 🏠 返回主页