cs.RO（2026-10-06）

📊 共 60 篇论文 | 🔗 8 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (40 🔗5) 支柱二：RL算法与架构 (RL & Architecture) (8 🔗1) 支柱九：具身大模型 (Embodied Foundation Models) (7 🔗1) 支柱七：动作重定向 (Motion Retargeting) (3) 支柱三：空间感知与语义 (Perception & Semantics) (2 🔗1)

🔬 支柱一：机器人控制 (Robot Control) (40 篇)

#题目一句话要点标签🔗⭐
1 From Legs to Wheels: Embodiment-Aware Human Motion Retargeting for Mobile-Base Humanoids 提出任务条件框架以解决移动底盘人形机器人运动重定向问题 legged robot humanoid motion retargeting
2 WareFly-VLA: A Vision-Language-Action Framework for UAV Navigation and Human Tracking in Smart Warehouses 提出WareFly-VLA框架以解决无人机在智能仓库中的语言引导导航问题 humanoid manipulation world model
3 Beyond Retargeting: Low-Latency and Robust Humanoid Whole-Body Teleoperation with Learned Atomic Motion Primitives 提出无重定向策略以解决人形机器人远程操作延迟与稳定性问题 humanoid teleoperation Unitree
4 EigenDEXplore: Structured Exploration for Dexterous Manipulation with Human Priors 提出EigenDEXplore以解决灵巧操作中的探索问题 manipulation dexterous hand dexterous manipulation
5 BiGym 2.0: Benchmarking Learned and Agent-Developed Policies for Humanoid Household Manipulation 提出BiGym 2.0以解决类人机器人家庭操作的挑战 humanoid whole-body control manipulation ✅
6 VOMMI: Collecting and Leveraging Portable Demonstrations for Mobile Manipulation 提出VOMMI以解决移动操控数据稀缺问题 manipulation mobile manipulation teleoperation
7 iGPC: Generative Motion Priors for Object-Aware Humanoid Interaction 提出iGPC框架以解决人形机器人与环境交互的挑战 humanoid humanoid robot humanoid control
8 Humanoid Horizon: Extending Task Horizon in Whole-Body Loco-Manipulation via Parallel Training, Dynamic Starting, and Reward Gating 提出Humanoid Horizon以解决长时间全身运动操控问题 humanoid humanoid robot manipulation
9 VLA-ACL: Action-Consistent Visual Token Pruning for Efficient Vision-Language-Action Models 提出VLA-ACL以解决视觉语言动作模型的高计算成本问题 manipulation vision-language-action VLA ✅
10 Model-Based Geometry-Aware Generative Optimization for Constrained Locomotion Planning 提出几何感知生成优化以解决受限运动规划问题 quadruped humanoid humanoid locomotion
11 DepthWorld: 3D World Model for Robot Manipulation 提出DepthWorld以解决机器人操作中的3D世界建模问题 manipulation world model world models
12 ESP: Energy-Score Policy for One-Step Multimodal Action Generation 提出ESP以解决多模态动作生成中的延迟问题 manipulation flow matching vision-language-action
13 Magnet-Aware Control of Legged Robots 提出磁场感知控制框架以解决腿式机器人在强磁场中的稳定性问题 quadruped legged robot whole-body control
14 SMART: Zero-Shot Sim-to-Real Articulated Object Manipulation via Large-Scale Synthetic Pretraining 提出SMART系统以解决关节物体操控的零样本仿真到现实问题 manipulation sim-to-real vision-language-action
15 QF3: Fast Flow RL with Filtered Q-Gradients 提出QF3以加速流政策的强化学习训练 humanoid humanoid locomotion motion tracking
16 ViDAL: A Visual Dynamics-Grounded Action Latent Space for Vision-Language-Action Models 提出ViDAL以解决视觉动态与动作表示不足的问题 dual-arm policy learning vision-language-action
17 PhoneBot: A Low-Cost Open Humanoid Robot Platform Reusing Smartphones 提出PhoneBot以解决人形机器人高成本和复杂性的挑战 humanoid humanoid robot locomotion
18 Feeling Through the Load: Compliant Quadruped Locomotion under Payload Interactions 提出一种力感知的四足机器人负载交互运动框架 quadruped locomotion
19 AutodidactWAM: Cross-Modal Self-Distillation from Generated Video to Robot Actions 提出AutodidactWAM以解决视频与机器人动作不匹配问题 humanoid teleoperation Unitree
20 Safe Multi-Robot Collaborative Transport Using Density Functions 提出基于密度函数的多机器人安全协作运输方法 quadruped manipulation MPC ✅
21 PEARS: Physical-Prior-Guided Efficient Adaptation via Failure Reasoning and Diffusion Steering for Tactile Manipulation 提出PEARS框架以解决机器人操作中的适应性问题 manipulation reinforcement learning flow matching ✅
22 Towards Efficient Robotic Manipulation Models with Self-Recursive Pruning 提出LCAM修剪方法以提高机器人操作模型的效率 manipulation OpenVLA
23 Commit While Futures Agree: Consequence-Aware Adaptive Action Chunking for Robot Manipulation 提出CA$^3$C以解决机器人操作中的行动承诺问题 manipulation world model world models
24 PACE: Stage-Consistent Long-Horizon Robot Manipulation via Progress-Aligned Context for Execution 提出PACE以解决长时间机器人操作中的阶段混淆问题 manipulation multimodal
25 Fast Non-Parametric Heteroscedastic Imitation Learning With Geometric Priors 提出非参数异方差模仿学习方法以解决数据稀缺问题 manipulation shared control imitation learning
26 Energy-Aware Path Following: Comparative Analysis of Reinforcement Learning and NMPC for Electric Vehicles 提出能量感知路径跟踪控制策略以优化电动车性能 model predictive control reinforcement learning PPO
27 Multi-Robot Multi-Goal Motion Planning with Stochastic Skills 提出基于随机技能的多机器人多目标运动规划方法 manipulation motion planning
28 TacZero: Training-Free Peg Insertion Using a General-Purpose Vision-Language Model with Tactile Feedback 提出TacZero以解决无训练的插入任务问题 manipulation motion estimation
29 Pareto-Optimal Entropy-Regularized Trajectory Optimization 提出Pareto最优熵正则化轨迹优化方法以解决非线性动态问题 trajectory optimization
30 Mitigating Concept Drift in QoS Prediction for Teleoperation of Autonomous Vehicles Using Historic Data 提出基于历史数据的预测框架以解决自主车辆远程操作中的QoS预测问题 teleoperation
31 IronMan: Information-Constrained Video-Action Learning for Robot Manipulation 提出IronMan框架以解决机器人操作中的视频动作生成问题 manipulation ✅
32 CoRE: Learning Collaboration-Role Experts for Decentralized Collaborative Manipulation with One Policy 提出CoRE以解决无任务提示的去中心化协作操控问题 manipulation
33 Micro Neural Policies for Safe Real-Time Robotic Control 提出微神经策略以解决资源受限机器人控制问题 sim-to-real zero-shot transfer
34 ExoBridge: Learning a Bare Hand to Hand-Worn Exoskeleton Mapping through Human Limb Coupling 提出ExoBridge以解决裸手与穿戴式外骨骼映射问题 manipulation bi-manual
35 Compact Robot Policies Need Fine-Grained Visual Representations 提出紧凑型机器人策略以解决视觉表示不足问题 manipulation flow matching
36 Post-Grasp Kinematic Repair for Robotic Insertion via Object-in-Gripper Reorientation 提出后抓取运动学修复方法以解决机器人插入问题 manipulation
37 Event-Driven Proactive Robot Assistance through Vision-Language Reasoning 提出事件驱动的主动机器人辅助方法以解决协作操作中的指令依赖问题 manipulation
38 Reactive Exploration of Unknown Environments for Redundant Robots using Virtual Model Control 提出主动探索方法以解决冗余机器人在未知环境中的导航问题 manipulation
39 OpenWAM: An Open Framework for Composable World-Action Models 提出OPENWAM框架以解决世界-动作模型设计比较难题 bi-manual
40 Beyond Task Reward: A Controller-Restriction Protocol for Evaluating Embodiment-Dependent Competence 提出控制器限制协议以评估机器人形态依赖能力 locomotion

🔬 支柱二：RL算法与架构 (RL & Architecture) (8 篇)

#题目一句话要点标签🔗⭐
41 StairVLA: Stage-Aware Hierarchical Action Generation for Vision-Language-Action Models 提出StairVLA以解决VLA模型生成动作效率低下问题 flow matching vision-language-action VLA
42 Modeling Latent Disturbances for Robust Decision-Making in World Models 提出潜在干扰模型以增强世界模型中的决策鲁棒性 world model world models latent dynamics
43 A Belief-State World Model for Catheter Navigation under Sparse Fluoroscopy: A Planar Proof of Concept 提出基于信念状态的世界模型以解决稀疏透视下的导管导航问题 world model world models
44 How Much Planning Is Enough? Reducing Search and Computation in World-Model Planning 提出SufficientPlan以解决世界模型规划中的搜索与计算效率问题 world model world models
45 RIWANav: Recursive World-Action Models with Self-Improvement for Urban Navigation 提出RIWANav以解决城市导航中的长期决策问题 reinforcement learning imitation learning world model
46 Sensor-Layout-Agnostic Navigation via Geometric Observation Canonicalization 提出一种传感器布局无关的导航方法以解决视觉导航的局限性 reinforcement learning zero-shot transfer
47 Seeing Through the Displaced Frame: Privileged Noise Distillation for Vision-Force Precision Assembly 提出特权噪声蒸馏以解决精密装配中的姿态误差问题 distillation ✅
48 Learning Grasp Targeting from Point Clouds for Log Pile Clearing on a Hydraulic Crane 提出基于点云学习抓取策略以解决液压起重机的木材堆清理问题 reinforcement learning behavior cloning

🔬 支柱九：具身大模型 (Embodied Foundation Models) (7 篇)

#题目一句话要点标签🔗⭐
49 Seeing the Invisible: Physics-Guided Visual Prompting for Temperature- and Radiation-Aware VLA Navigation 提出物理引导视觉提示以解决温度和辐射风险导航问题 vision-language-action VLA VLN
50 ActTune: Action-Aware Precision and GPU Operating-Point Adaptation for Energy-Efficient Vision-Language-Action Inference 提出ActTune以解决GPU能耗与任务成功率的平衡问题 vision-language-action VLA
51 MIM-VLA: Learning Physical Interaction Representations from Gripper Motor Feedback 提出MIM-VLA以解决视觉-语言-动作中的物理交互表示问题 vision-language-action VLA
52 Adapting Vision-Language-Action Models to Unknown Visual Disruptions During Execution 提出SALT以应对执行过程中的未知视觉干扰问题 vision-language-action VLA
53 A Swarm-Coordinated Multi-Robot System for Early Stress Detection in Agricultural Rows Using Multimodal Leaf Sensing 提出CropSentry以解决农业作物早期压力检测问题 multimodal
54 OntoPlan: An Ontology-Grounded Scene Representation and Agentic Framework for Scalable Robot Task Planning 提出OntoPlan以解决机器人任务规划中的空间信息捕捉问题 large language model instruction following ✅
55 Navigation with RF Cues: Embodied Perception Action under Multipath Uncertainty 提出多模态导航框架以解决RF信号多径不确定性问题 multimodal

🔬 支柱七：动作重定向 (Motion Retargeting) (3 篇)

#题目一句话要点标签🔗⭐
56 One for All, All for One: Coordinated Multi-Agent Diffusion Steering via Stochastic Optimal Control 提出协调多智能体扩散引导方法以解决生成模型的结构化输出问题 human motion
57 Silicon Language: A Robot-Native Knowledge Exchange Framework for Heterogeneous Robots 提出Silicon Language框架以解决异构机器人知识共享问题 cross-embodiment
58 The Robot Is Not Its Description: GaugeBench for Representation Robustness in Morphology-Aware Policies 提出GaugeBench以解决机器人描述与行为不一致问题 cross-embodiment

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

#题目一句话要点标签🔗⭐
59 OpenSplatGraph: From Dense Semantic Maps to Structured Scene Graphs for Open-Vocabulary Robot Perception 提出OpenSplatGraph以解决机器人感知中的语义映射问题 3D gaussian splatting gaussian splatting splatting ✅
60 EgoLAP: Learning from Egocentric Human Data through Language-Action Reasoning 提出EgoLAP以解决机器人学习中的人类数据利用问题 affordance egocentric VLA

⬅️ 返回 cs.RO 首页 · 🏠 返回主页