cs.RO（2026-05-28）

📊 共 31 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (25 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (4 🔗2) 支柱三：空间感知与语义 (Perception & Semantics) (2)

🔬 支柱一：机器人控制 (Robot Control) (25 篇)

#题目一句话要点标签🔗⭐
1 Qwen-VLA: Unifying Vision-Language-Action Modeling across Tasks, Environments, and Robot Embodiments Qwen-VLA：统一视觉-语言-动作建模，实现跨任务、环境和机器人形态的通用具身智能 manipulation egocentric vision-language-action
2 BORA: Bridging Offline Reinforcement Learning and Online Residual Adaptation for Real-World Dexterous VLA Models BORA：桥接离线强化学习与在线残差自适应，用于真实世界灵巧VLA模型 manipulation dexterous hand dexterous manipulation
3 Gaze2Act: Gaze-Conditioned Vision-Language-Action Policies for Interactive Robot Manipulation Gaze2Act：利用注视引导的视觉-语言-动作策略实现交互式机器人操作 humanoid manipulation Unitree
4 VLA-Pro: Cross-Task Procedural Memory Transfer for Vision-Language-Action Models VLA-Pro：面向视觉-语言-动作模型的跨任务程序记忆迁移框架 manipulation vision-language-action VLA
5 VE2VF: Vision-Enabled to Vision-Free Distillation via Real-world Reinforcement Learning for Robust Contact-Rich Manipulation 提出VE2VF，通过真实世界强化学习蒸馏实现接触式操作的鲁棒性 manipulation domain randomization reinforcement learning
6 3DVLA: Enhancing Vision-Language-Action Models via 3D Spatial and Instance Understanding 提出3DVLA框架，通过3D空间和实例理解增强视觉-语言-动作模型 manipulation scene understanding vision-language-action
7 Phase-Conditioned Imitation Learning with Autonomous Failure Recovery for Robust Deformable Object Manipulation 提出基于阶段条件模仿学习与自主故障恢复的柔性物体操作框架 manipulation dual-arm teleoperation ✅
8 PhAIL: A Real-Robot VLA Benchmark and Distributional Methodology PhAIL：基于Franka FR3真实机器人的VLA基准测试与分布评估方法 teleoperation vision-language-action VLA
9 VLAConf: Calibrated Task-Success Confidence for Vision-Language-Action Models 提出VLAConf，用于校准视觉-语言-动作模型的任务成功置信度，提升机器人操作的可靠性。 manipulation vision-language-action VLA
10 ElegantVLA: Learning When to Think for Efficient Vision-Language-Action Models ElegantVLA：通过学习何时思考，加速高效的视觉-语言-动作模型 manipulation vision-language-action VLA
11 MARS Policy: Multimodality Only When It Matters 提出MARS策略，自适应地在机器人操作中引入多模态，提升效率。 manipulation imitation learning multimodal
12 Sample-Efficient Diffusion-based Reinforcement Learning with Critic Guidance 提出CGPO：一种基于Critic引导的扩散强化学习方法，提升样本效率。 locomotion reinforcement learning diffusion policy
13 MonoDuo: Using One Robot Arm to Learn Bimanual Policies MonoDuo：利用单臂机器人学习双臂操作策略，解决双臂机器人数据稀缺问题 manipulation bi-manual bimanual manipulation
14 Decentralized LLM-Driven Coordination of Acoustic Robots for Contactless Object Manipulation 提出基于LLM的分布式声学机器人协同框架，实现非接触式物体操作 manipulation large language model
15 DynaFLIP: Rethinking Robotics Perception via Tri-Modal-Dynamics Guided Representation 提出DynaFLIP以解决机器人感知中的动态理解问题 manipulation VLA multimodal
16 LLM-Guided Future Hypotheses for Horizon-Aware Exploration in Multi-Step Robot Manipulation 提出基于LLM引导的未来假设方法，用于多步机器人操作中的前瞻性探索。 manipulation reinforcement learning
17 RoboWits: Unexpected Challenges for Robotic Creative Problem Solving RoboWits：提出双臂机器人创造性问题解决基准，应对意外挑战。 manipulation bi-manual VLA ✅
18 A Progress-Aware Leader-Follower Midair Docking System for Dual-Drone Aerial Manipulation 提出一种基于进度感知的双无人机空中对接系统，用于空中操作 manipulation
19 The Open Motion Planning Library 2.0 OMPL 2.0：面向实时运动规划的硬件加速开源库 motion planning
20 Extreme dynamic symmetry enables omnidirectional and multifunctional robots 提出动态对称性概念，设计全向多功能机器人Argus，提升敏捷性与鲁棒性。 locomotion
21 ELAN4D: Embodiment-Centric 4D Supervision for Vision-Language-Action Models via Plug-and-Play Adaptation ELAN4D：通过即插即用适配，实现以具身认知为中心的VLA模型4D监督 manipulation vision-language-action VLA
22 Physics-informed Goal-Conditioned Reinforcement Learning under Hybrid Contact Dynamics 提出接触感知的分层物理信息强化学习，解决接触动力学下的目标条件强化学习问题 manipulation reinforcement learning contact-aware
23 Any-ttach: Quick End-effector Swapping Enables Manipulation Dexterity with Simplicity Any-ttach：通过快速末端执行器切换实现灵巧操作，简化机器人操作复杂性 manipulation ✅
24 ARISTO Hand: Sensing-Driven Distal Hyperextension for Fine-Grained Manipulation ARISTO Hand：通过力觉驱动远端过伸实现精细操作 manipulation
25 VLM-GLoc: Vision-Language Model Enhanced Monte Carlo Localization for Robust Semantic Global Localization in Cluttered Quasi-Static Environments VLM-GLoc：利用视觉-语言模型增强蒙特卡洛定位，实现复杂准静态环境下的鲁棒语义全局定位 quadruped open-vocabulary open vocabulary

🔬 支柱二：RL算法与架构 (RL & Architecture) (4 篇)

#题目一句话要点标签🔗⭐
26 A Heterogeneous Architecture for Robot RL Beyond GPU-Dominant Paradigms 提出UniLab异构架构，解耦CPU仿真与GPU学习，提升机器人强化学习训练效率。 policy learning PPO SAC ✅
27 ZAPS-DA: Zero-Phase Action Policy Smoothing with Decoupled Actor for Continuous Control in Reinforcement Learning ZAPS-DA：零相位动作策略平滑与解耦Actor，解决强化学习连续控制中的抖动问题 reinforcement learning distillation
28 Learning-Based Navigation for Indoor Mobile Robots 提出一种基于学习的室内移动机器人导航框架，融合全局规划和局部控制。 reinforcement learning PPO behavior cloning ✅
29 UniLab: A Heterogeneous Architecture for Robot RL Beyond GPU-Dominant Paradigms 提出UniLab以解决机器人强化学习中的GPU依赖问题 policy learning PPO

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

#题目一句话要点标签🔗⭐
30 EXACT-MPPI: Exact Signed-Distance Navigation for Arbitrary-Footprint Robots from Point Clouds via Path Integral Control 提出EXACT-MPPI以解决复杂足迹机器人导航问题 occupancy grid
31 CoMo3R-SLAM: Collaborative Monocular Dense SLAM with Learned 3D Reconstruction Priors for Outdoor Multi-Agent Systems CoMo3R-SLAM：面向户外多智能体系统的协同单目稠密SLAM，利用学习的3D重建先验 3D reconstruction feature matching

⬅️ 返回 cs.RO 首页 · 🏠 返回主页