cs.RO（2026-09-28）

📊 共 67 篇论文 | 🔗 14 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (44 🔗9) 支柱二：RL算法与架构 (RL & Architecture) (11 🔗2) 支柱九：具身大模型 (Embodied Foundation Models) (7 🔗3) 支柱三：空间感知与语义 (Perception & Semantics) (2) 支柱六：视频提取与匹配 (Video Extraction) (2) 支柱四：生成式动作 (Generative Motion) (1)

🔬 支柱一：机器人控制 (Robot Control) (44 篇)

#题目一句话要点标签🔗⭐
1 DexRoam: Learning Mobile Bimanual Dexterous Manipulation from Egocentric Whole-Body Human Demonstrations 提出DexRoam以解决移动双手灵巧操作的示范瓶颈问题 locomotion manipulation mobile manipulation
2 DexWeave: Learning Dexterous Humanoid Loco-Manipulation from Human Demonstrations 提出DexWeave以解决人类示范下的灵巧人形机器人运动问题 humanoid manipulation loco-manipulation
3 Uni-VLaT: Whole-Body Tactile Adaptation of VLA Policies for Humanoid Loco-Manipulation 提出Uni-VLaT以解决类人机器人在运动操作中的触觉适应问题 humanoid locomotion manipulation
4 Humanoid Loco-Manipulation With Discrete VLA Model 提出Holo-M以解决类人机器人运动与操作的高维控制问题 humanoid manipulation loco-manipulation
5 WB-WAM: Heterogeneous Body-Hand Pre-training for Humanoid Loco-Manipulation 提出WB-WAM以解决人形机器人运动与操作协调问题 humanoid manipulation loco-manipulation
6 Model-Informed Safe Reinforcement Learning for Bipedal Locomotion via Step-to-Step Prediction 提出基于模型的安全强化学习框架以解决双足机器人步态控制问题 humanoid humanoid robot bipedal
7 CollisionSplatting: Collision-Aware Motion Planning in 3DGS Scenes with Image-Conditioned Objectives and Adjustable Conservatism 提出CollisionSplatting以解决3D场景中的运动规划问题 manipulation motion planning 3D gaussian splatting
8 Gaze Prompts: Temporally Dense Human Attention for Vision-Language-Action Fine-Tuning 提出眼动追踪引导以提升视觉-语言-动作微调效果 manipulation bi-manual bimanual manipulation
9 HOI-Retarget: Contact-Centric Retargeting for Human-Object Interaction 提出HOI-Retarget以解决人机交互中的运动数据生成问题 humanoid humanoid robot trajectory optimization
10 GAE: General Action Expert for Real-Time Humanoid Teleoperation 提出GAE框架以解决实时类人机器人远程操作中的同步问题 humanoid domain randomization teleoperation ✅
11 Unified Visual-Tactile-Action Modeling from Human Demonstrations for Dexterous Manipulation 提出统一视觉-触觉-动作模型以解决灵巧操作中的触觉反馈问题 manipulation dexterous hand dexterous manipulation ✅
12 Learning to Act under Visual Interruptions with Vision-Language-Action Models 提出MAIL-Bench和MINT以解决视觉中断下的机器人操作问题 manipulation world model world models ✅
13 MM-ABC: Towards Generalist Mobile Manipulation via Seeing, Coordinating and Imagining 提出MM-ABC以解决移动操控中的感知与协调问题 manipulation mobile manipulation world model
14 DexAgent: An Agentic Human2Sim2Robot Framework for Dexterous Manipulation with Self-Evolving Tool Library 提出DexAgent框架以解决灵巧机器人操作中的多样性问题 manipulation dexterous manipulation sim-to-real ✅
15 Do Not Cut When Uncertain: Rejectable and Calibrated Decision Heads for VLA Policies in Robotic Harvesting 提出可拒绝和校准决策头以解决机器人采摘中的不确定性问题 manipulation behavior cloning flow matching
16 RoboIRGBench: Benchmarking Implicit Referential Grounding in Vision-Language-Action Models 提出RoboIRG-Bench以解决隐式指称基础的机器人任务评估问题 manipulation vision-language-action VLA
17 Spatial Grafting: Grounding 3D Features for Flow-Matching Robot Policies 提出空间嫁接方法以解决机器人策略中的几何信息缺失问题 manipulation mobile manipulation dual-arm
18 Proprioceptive Force Estimation for Quadruped Locomotion and Human-Robot Interaction 提出一种新方法以实现四足机器人运动中的自我感知力估计 quadruped locomotion locomotion policy
19 Rethinking Causal Action Tokenization with Conditional Annealing in Flow Matching 提出CATok以解决现有动作标记不一致问题 manipulation flow matching vision-language-action
20 WAM-OPD: Sharpening World Action Models via On-Policy Distillation 提出WAM-OPD以解决机器人任务适应性问题 manipulation world action model world action models
21 Where Memory Belongs: Ledger, an Object Ledger for Memory-Augmented VLAs 提出Ledger以解决长时记忆不足的问题 manipulation vision-language-action VLA
22 RAVEL: Asynchronous Rolling Inference for Flow-Based Vision-Language-Action Models 提出RAVEL以解决流式视觉-语言-动作模型的延迟问题 manipulation vision-language-action VLA
23 mmHRI: Towards Privacy-Preserving Human-Robot Interaction with Millimeter-Wave Radar 提出mmHRI以解决隐私保护的人机交互问题 manipulation human motion vision-language-action
24 Terrain-Aware Autonomous Planetary Exploration for Exteroceptive-Proprioceptive Mapping with Quadruped Scouts 提出一种地形感知的自主行星探索框架以解决未知地形导航问题 quadruped locomotion elevation map
25 From Pixel to Poses: Object-centric Tool Manipulation Learning from Human Demonstrations 提出P2P-T框架以解决机器人工具操作数据稀缺问题 manipulation world model world models
26 ReCAT: Remember, Count, and Time: Structured Recurrent Memory for Robot Manipulation 提出ReCAT以解决机器人操控中的记忆依赖问题 manipulation flow matching Mamba ✅
27 QuadHand: A Compact Quadrotor Aerial Manipulator with MRC-SDF-Based Whole-Body Motion Planning 提出QuadHand以解决无人机操作中的干扰问题 manipulation trajectory optimization motion planning
28 CoHuB: A Simulation Benchmark for Multi-Humanoid Collaboration 提出CoHuB以解决多类人形机器人协作评估问题 humanoid teleoperation VR teleoperation
29 UMR: Universal Manipulation Representation 提出UMR以解决跨机器人操作技能转移问题 manipulation cross-embodiment VLA ✅
30 Self-Evolving Coding Agents: From Digital Programs to Physical-World Intelligence 提出物理编码方法以解决机器人执行中的泛化问题 dual-arm vision-language-action VLA
31 EMPIRIC: Experiment-Driven Learning of Residual World Models for Robot Planning 提出EMPIRIC以解决机器人规划中的未知机制学习问题 manipulation world model world models ✅
32 Don't Throw Away the Tail: Action Upcycling for Policy Acceleration 提出动作再利用以加速机器人策略执行 manipulation world action model world action models
33 Sufficiency of Zeroth-Order Reward Shaping for Policy Gradient in Stabilization Control 提出零阶奖励塑形以解决稳定控制中的策略梯度问题 trajectory optimization reinforcement learning deep reinforcement learning
34 TLC-DiT: Task-Aligned Local Visual Conditioning for Robust Multitask Robot Manipulation 提出TLC-DiT以解决多任务机器人操作中的视觉信息不足问题 manipulation bi-manual language conditioned
35 F4R: Failure-Driven Recognition, Reconstruction, Refinement, and Redeployment for Continual Robot Self-Improvement 提出F4R框架以解决机器人自我改进中的失败识别与学习问题 manipulation sim-to-real reinforcement learning
36 On the Numerical Reliability of Differentiable Physics-Based Optimization for Robotic Material Manipulation 研究可微物理优化在机器人材料操作中的数值可靠性问题 manipulation differentiable simulation
37 LQR-ArUco Fusion: Robust Hierarchical Control for Navigation and Asymmetric Manipulation in Two-Wheeled Robots 提出层次控制框架以解决双轮机器人导航与不对称操作问题 manipulation
38 Graph-Based Simultaneous Path and Foothold Planning for Multi-Limbed Intra-Vehicular Robots in Space Stations 提出基于图论的路径与支撑点规划框架以解决多肢体机器人问题 locomotion manipulation motion planning
39 RoboICL: Embodied In-Context Learning with GPT-6 Astra 提出RoboICL框架以提升机器人控制精度与效率 manipulation VLA ✅
40 Reliability-Aware Sparse Route Memory for Round-Trip Vision-Language Navigation 提出可靠性感知稀疏路线记忆以解决往返视觉语言导航问题 Unitree VLN
41 ARS: Agentic Reward System for Robot Learning 提出Agentic Reward System以解决机器人学习中的进度奖励建模问题 manipulation policy learning ✅
42 RoboCompiler: Graph-Native Compilation of Closed-Chain Robots for Consistent Modeling, Control, and Simulation 提出RoboCompiler以解决闭链机器人建模与控制一致性问题 Unitree
43 From Language to Task Maps: Compiling Semantic Relations While Preserving Task-Relevant Freedom 提出语义到任务地图的编译方法以解决机器人控制中的约束问题 manipulation
44 Predictive Semantic Safety: From Visual Physical Reasoning to Safety-Critical Control 提出预测语义安全框架解决机器人安全控制问题 Unitree

🔬 支柱二：RL算法与架构 (RL & Architecture) (11 篇)

#题目一句话要点标签🔗⭐
45 From World Models to World Action Models: Rethinking Next-State Prediction 提出CF-WAM以解决动态状态预测的限制问题 world model world models world action model
46 RoboFL: Federated Expert Assembly for World Action Models 提出RoboFL以解决世界行动模型中的数据稀缺问题 world action model world action models distillation
47 AD-E2E-JEPA: A Joint-Embedding Predictive Architecture For End-to-End Autonomous Driving 提出AD-E2E-JEPA以解决端到端自动驾驶中的世界模型效率问题 policy learning imitation learning world model ✅
48 Natural State-Prediction Accuracy can Hide Weak Controlled Responsiveness in VLA Readouts 提出评估框架以解决VLA模型响应性不足问题 MAE vision-language-action VLA
49 Quantile Head for Vision-Language-Action Models 提出量化头以解决视觉-语言-动作模型的动作估计问题 flow matching vision-language-action VLA ✅
50 Efficient World Action Model Inference with Adaptive Intermediate States 提出WAMACHINE以解决世界动作模型推理效率问题 world action model world action models
51 FutureDuet: Decoupling Observation Access from Future Supervision in World Action Models 提出FutureDuet以解决WAMs中观察访问与未来监督的耦合问题 world action model world action models
52 When World Models Lie: Adaptive Safety Analysis Under Wrong Imaginations 提出自适应安全分析方法以应对世界模型错误问题 world model world models
53 Zero-Shot Reactive Obstacle Avoidance for Generative Robot Policies 提出NUDGE以解决机器人政策中的障碍物规避问题 diffusion policy flow matching vision-language-action
54 Agent Priors-guided Policy Learning 提出基于代理先验的策略学习以解决技能组合与泛化问题 policy learning
55 RLE-Bench: A Qualifying Exam for Coding Agents as Robot Learning Engineers 提出RLE-Bench以评估机器人学习工程师的编码能力 policy learning multimodal

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

#题目一句话要点标签🔗⭐
56 EdgeVLN: Runtime-Aware Deployment Ready Quantized Vision Language Navigation Model 提出EdgeVLN以解决边缘设备上VLN模型部署问题 vision-language-navigation VLN
57 FailPatch: Failure Residual Patching for Vision-Language-Action Models 提出FailPatch以解决视觉-语言-动作模型中的失败状态问题 vision-language-action VLA ✅
58 NavJev: Efficient Vision-Language Navigation via Action-Centric Visual Compression and Discriminative Action-Semantic Memory 提出NavJev以解决视觉语言导航中的推理延迟问题 VLN large language model multimodal ✅
59 Denoising Multi-Robot Trajectories 提出D4orm框架以解决多机器人轨迹规划问题 multimodal ✅
60 NavHarness: Towards Lifelong Embodied Navigation 提出NavHarness以解决长期嵌入式导航中的记忆处理问题 multimodal
61 Bayesian Active Learning for Intent Disambiguation in Interactive Robot Planning 提出贝叶斯主动学习框架以解决机器人规划中的意图歧义问题 large language model
62 SAGE: Symbolic Action-Gating and Editing for LLM Task Planners 提出SAGE以解决LLM任务规划中的安全性与效率问题 large language model

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

#题目一句话要点标签🔗⭐
63 ForVis: An In-Field Dataset and Benchmark for VIO Using Under-Canopy UAV Flights in Forests 提出ForVis数据集以解决森林环境下VIO评估问题 VIO
64 ExcavaTwin: Training-Free Geometry-Guided Semantic Elevation Mapping for Autonomous Excavation 提出ExcavaTwin以解决自主挖掘中的几何与语义映射问题 elevation map

🔬 支柱六：视频提取与匹配 (Video Extraction) (2 篇)

#题目一句话要点标签🔗⭐
65 MonoEgo: Monocular Metric Egocentric Demonstration Capture with Passive Wrist Constellations and Sparse Workstation Anchors 提出MonoEgo以解决单目度量演示捕获问题 egocentric
66 Action Sequence Transfer via LLMs for Heterogeneous Environments 提出动作序列迁移系统以解决异构环境中的用户行为适应问题 Ego4D

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
67 Contact-Aware Impedance Controller for Robot-Assisted Ultrasound Imaging 提出基于B模式超声图像的接触感知控制器以提升机器人超声成像安全性 contact-aware

⬅️ 返回 cs.RO 首页 · 🏠 返回主页