cs.RO（2026-10-08）

📊 共 82 篇论文 | 🔗 15 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (62 🔗11) 支柱二：RL算法与架构 (RL & Architecture) (10 🔗3) 支柱三：空间感知与语义 (Perception & Semantics) (5) 支柱九：具身大模型 (Embodied Foundation Models) (3) 支柱七：动作重定向 (Motion Retargeting) (1) 支柱六：视频提取与匹配 (Video Extraction) (1 🔗1)

🔬 支柱一：机器人控制 (Robot Control) (62 篇)

#题目一句话要点标签🔗⭐
1 Generative Neural Retargeting for Human-to-Robot Dexterous Manipulation 提出生成神经重定向方法以解决人机灵巧操作问题 manipulation dexterous manipulation MPC
2 SimVLA: Zero-Shot Sim-to-Real VLA Learning for Mobile Manipulation 提出SimVLA以解决移动操控中的真实数据采集难题 manipulation mobile manipulation sim-to-real
3 A Physics-Informed Collision Learning Framework for Collaborative Robot Motion Generation 提出PI-UDF框架以解决协作机器人运动生成中的碰撞预测问题 manipulation dual-arm MPC
4 Humanoid World Action Model With Joint State--Action Generation 提出HWAM以解决人形机器人动作执行差异问题 humanoid humanoid robot whole-body control
5 DAMP: Humanoid Locomotion via Denoised Belief Learning and Adversarial Motion Priors 提出DAMP框架以解决类人机器人在复杂环境中的稳定行走问题 humanoid humanoid robot humanoid locomotion
6 Being-M0.7: A Latent World-Action Model for Humanoid Robots 提出Being-M0.7以解决人形机器人运动与操作协调问题 humanoid humanoid robot humanoid control
7 LeWAM: A JEPA World Action Model with Diffusion-Steering-Based MPC 提出LeWAM以解决世界动作模型中的噪声冗余问题 MPC flow matching world model
8 PathTime-VLA: Path-Time Decoupling for Factorized Post-Training of Vision-Language-Action Policies 提出PathTime-VLA以解决视觉-语言-动作策略的时间耦合问题 motion planning teleoperation vision-language-action
9 VioLA: Learning Generalist Humanoid Control Policies from Human Data 提出VioLA以解决人形机器人控制策略学习问题 humanoid humanoid control locomotion
10 WARP-VLA: Wrist-Camera Adaptation for View-Robust Policy Execution in Vision-Language-Action Models 提出WARP-VLA以解决腕部摄像头配置变化带来的挑战 manipulation vision-language-action VLA
11 REACT: Rolling Denoising and Dual Decoupling for Reactive Robot Control with VLA Models 提出REACT框架以提升机器人控制的反应性和流畅性 manipulation bi-manual bimanual manipulation
12 Dex-One2Many: Learning Dexterous Manipulation from a Single Human Demonstration 提出Dex-One2Many以解决单一人类示范学习的泛化与探索问题 manipulation dexterous manipulation sim-to-real
13 UNITAS: A 3D-Native World Action Model for Embodied Manipulation 提出UNITAS以解决三维空间机器人操控中的世界动作建模问题 manipulation world action model world action models ✅
14 SpatialHarness: Test-Time Spatial Scaffolding for Fine Robotic Manipulation 提出SpatialHarness以解决精细机器人操作中的空间可观察性问题 manipulation spatial relationship foundation model ✅
15 A Balanced Data Diet: Addressing the Exploration Bottleneck in Mega-Scale RL for Robot Control 提出成功引导采样以解决大规模强化学习中的探索瓶颈问题 quadruped locomotion manipulation ✅
16 Toward Lunar Legged Robots: Field Deployment Lessons at LUNA 提出针对月球表面的四足机器人行走策略以解决环境挑战 quadruped legged robot locomotion
17 ManiUnit: A Manipulation Skill Dataset and Benchmark for Long-Horizon Tasks 提出ManiUnit以解决长时间移动操控任务中的技能评估问题 manipulation mobile manipulation vision-language-action
18 From Solo to Ensemble: A Hierarchical Framework for Composable Multi-Agent Human-Object Interaction 提出层次化框架以解决多智能体人机交互问题 manipulation policy learning human-object interaction
19 A Minimal Optical-Flow Representation for Vision-Based Tactile Rotation Classification in Robotic Manipulation Across Gravity Domains 提出一种紧凑的光流表示法以解决机器人操控中的旋转分类问题 manipulation optical flow
20 OmniDex: Scaling Dexterous Hand Grasping to Diverse Cluttered Scenes 提出OmniDex以解决复杂场景下灵巧抓取问题 dexterous hand embodied AI multimodal
21 SkillWeave: Weaving Heterogeneous Demonstrations into Long-Horizon Manipulation Skills 提出SkillWeave框架以解决长时间操作技能的演示收集问题 manipulation dexterous manipulation teleoperation
22 Tell Robot What Not to Do: A Negation Understanding Perspective 提出NegaAlign以解决机器人遵循否定指令的问题 manipulation vision-language-action VLA
23 Acting from Belief, Looking When Needed: A Bayesian Spatial World Model for Navigation under Intermittent Perception 提出ALONE模型以解决间歇感知下的机器人导航问题 motion planning world model world models
24 Rewiring Semantics, Dynamics, and Control: A Simple yet Effective Action-Centric Tri-Stream Transformer 提出ACT³以解决机器人控制中的语义与动态信息融合问题 manipulation world model world models
25 WAM-Cache: Staleness-Bounded KV Reuse for Efficient World Action Models 提出WAM-Cache以解决视频DiT预填充计算成本问题 manipulation world action model world action models
26 RESETTLE: Robotic Recovery through Disagreement-Triggered Retrieval and Efficient Corrective Control 提出RESETTLE以解决机器人恢复效率低下的问题 manipulation trajectory optimization JEPA ✅
27 SuperNav: An Agentic Navigation System for Any Task in Any Scene 提出SuperNav以解决服务机器人导航任务的通用性问题 quadruped large language model multimodal ✅
28 YOCO: You Only Calibrate Once! Fast Mocap Calibration for Dexterous Teleoperation 提出YOCO以解决低成本动作捕捉手套的校准问题 teleoperation MANO
29 RoboAware: Learning to Coordinate Embodied Skills from Counterfactual Outcomes 提出RoboAware以解决机器人技能协调问题 bi-manual VLA Code as Policies
30 Experience-Guided Initiation Search for Learned Skills in Skill Composition 提出EVIS框架以解决新环境中技能组合的初始化配置问题 manipulation vision-language-action VLA
31 Walking on Roofs: Exploring the Potential of Walking Robots for Construction Work on Roofs 探索四足机器人在屋顶施工中的应用潜力 quadruped locomotion Unitree
32 Real-Time Motion Planning with Dynamic Hazards: Classical vs. Learning-Based Methods 通过统一基准比较经典与学习方法解决动态危险场景中的实时运动规划问题 motion planning PPO
33 Sim-to-Real RL for ASVs using SysID 提出一种新型ASV模拟器以解决动态环境中的控制策略问题 sim-to-real reinforcement learning
34 TACROSS: An Efficient and Low-Cost Scalable Human Touch System Across Heterogeneous Tactile Sensors for Dexterous Robot Learning 提出TACROSS以解决异构触觉传感器间数据对齐问题 manipulation teleoperation policy learning ✅
35 Unifying Policy Learning and State Prediction through Spatial Language Modeling 提出空间语言建模以统一策略学习与状态预测 manipulation policy learning
36 PMTRM: Pseudo-Memory Temporal Re-encoding Module for Embodied Policy Learning 提出伪记忆时间重编码模块以解决机器人操作中的阶段模糊问题 manipulation policy learning
37 MAP2: Model- and Acceleration-Based Pursuit with MPC and Gaussian Process Residual Learning for Autonomous Racing 提出MAP2以解决自主赛车中的轨迹跟踪问题 MPC
38 PIER: An Evidence-Gated Execution Interface for Robotic Manipulation 提出PIER以解决机器人操作中的执行授权问题 manipulation
39 Reliability-Aware Future Conditioning for Temporally Robust Robot Manipulation 提出可靠性感知未来条件以解决机器人操作的时间鲁棒性问题 manipulation ✅
40 FAITH: Feasibility-Aware Safety-Filtered RL for High-Dimensional Systems 提出FAITH框架以解决高维系统的安全强化学习问题 humanoid Unitree reinforcement learning
41 Control-Ready Uncertainty for Trajectory Diffusion 提出SCOPE以解决扩散模型不确定性评估问题 manipulation multimodal ✅
42 Embodied Turing Machines: Stateful Code for Robot Recursive Self-Improvement 提出COAP以解决机器人自我改进中的决策问题 bi-manual VLA
43 FOCUS: From Privileged States to RGB-D with Controlled Modality Switching and Representation Alignment 提出FOCUS以解决视觉强化学习中的模态差异问题 manipulation reinforcement learning PPO
44 RAGNAROK: Radar-Aided Gravity-Normalized Alignment for Robust Open Keyframe-based Radar-Visual-Kinematic-Inertial SLAM 提出RAGNAROK以解决复杂环境下的SLAM问题 legged robot contact-aware ✅
45 USDCraft: Geometrically Grounded Programmatic Modeling of Articulated 3D Assets for Simulation 提出USDCraft以解决真实与仿真机器人操作中的3D资产建模问题 manipulation sim-to-real
46 Towards Path-Creative Navigation: Robot Navigation through Embodied Interaction 提出路径创造导航以解决复杂环境中的自主导航问题 locomotion traversability
47 CSF: Contextual Safety Filtering for Motion Generators 提出上下文安全过滤器以解决运动生成器的安全性问题 Unitree
48 Traceable World State: A Provenance-Aware State Representation and Deterministic Replay Framework for Robotic Systems 提出可追溯世界状态框架以解决机器人系统状态表示问题 Unitree
49 Autonomous thermodynamic cycles via robotic mobility and sensing 提出自主热力循环以解决能量转换效率问题 locomotion
50 PriMoRetarget: Humanoid Motion Transfer Through Prioritized Motion Retargeting 提出PriMoRetarget以解决多角色运动重定向问题 humanoid humanoid robot manipulation
51 PanTact: Teaching Whole-Body Dexterity with a Wearable Contact Interface 提出PanTact以解决人形机器人全身灵巧性问题 humanoid humanoid robot whole-body control
52 Reflex: Visually Grounded Reactive Humanoid Control 提出Reflex框架以解决类人机器人视觉反应控制问题 humanoid humanoid control whole-body control
53 Spa-Bench: A Real-Robot Benchmark for Spatial Grounding in Vision-Language-Action Models 提出Spa-Bench以解决视觉-语言-动作模型空间定位问题 manipulation vision-language-action VLA ✅
54 StereoEngine: Scaling Stereo Bimanual Robot Data without Stereo Cameras 提出StereoEngine以解决机器人操作中缺乏立体数据的问题 manipulation bi-manual bimanual manipulation
55 Gamma: Grounded Multi-Agent Memory with Mechanical Audit for Memory-Intensive Manipulation 提出GAMMA以解决记忆密集型操作中的记忆映射问题 manipulation vision-language-action VLA
56 Graph-Based Heuristics for Minimum Constraint Removal in Cluttered Robot Motion Planning 提出图基启发式方法以解决机器人运动规划中的最小约束移除问题 humanoid manipulation motion planning
57 REMM: Revisable Embodied Memory for Manipulation 提出REMM以解决长时间操作中的记忆修正问题 manipulation vision-language-action ✅
58 CASTER: Formalizing Zero-Shot Manipulation from Synthetic Videos as Constrained Optimization 提出CASTER以解决从合成视频到零-shot操作的优化问题 manipulation
59 Fly-by-Code: Embodied Coding Agents for Aerial Manipulation with Active Visual and Physical Feedback 提出主动反馈框架以解决无人机操作中的反馈不足问题 manipulation
60 Tactile Sensing-based Secure and Gentle Grasping of Unknown Objects 提出基于触觉传感的安全柔性抓取方法以解决未知物体操控问题 manipulation model predictive control
61 Towards Path-Creative Navigation: Robot Navigation through Embodied Interaction 提出路径创造导航以解决复杂环境中的自主导航问题 locomotion traversability
62 Non-Prehensile Multi-Object Placement with Mobile Robots via Diffusion-Based Trajectory Planning 提出扩散规划方法以解决移动机器人多物体精确放置问题 manipulation diffusion policy

🔬 支柱二：RL算法与架构 (RL & Architecture) (10 篇)

#题目一句话要点标签🔗⭐
63 PLaW-VLA: Predictive Latent World Modeling for Vision-Language-Action Policies 提出PLaW-VLA以解决长时间控制中的世界预测问题 world model world models vision-language-action
64 Learning Language-Conditioned Traversability Representations for Adaptive Visual Navigation 提出LaTraNav框架以解决视觉导航中的可通行性问题 flow matching traversability language conditioned
65 DreamTrue: Action-Faithful Robot World Model with Counterfactual Post-Training 提出DreamTrue以解决机器人世界模型的动作跟随与交互覆盖问题 reinforcement learning world model world models ✅
66 MiniWAM: Learning Compact Future Targets for Efficient World-Action Modeling 提出MiniWAM以解决高维目标预测带来的训练成本问题 policy learning world model world models ✅
67 LiteNWM: Efficient Latent World Models for Onboard Visual Navigation in the Wild 提出LiteNWM以解决视觉导航中的未来预测问题 world model world models
68 Residual Modeling Closes the Regression and Generative Policy Gap in Robot Learning 提出异方差学生t动作回归以缩小机器人学习中的回归与生成策略差距 policy learning flow matching vision-language-action ✅
69 CAPABLE: Capability-Aware Policy Adaptation via Behavioral Latent Encoding 提出CAPABLE框架以解决视觉-语言-动作策略的适应性问题 reinforcement learning vision-language-action VLA
70 PlanWAM: Planning-Shaped Future Representations for End-to-End Autonomous Driving 提出PlanWAM以优化自主驾驶中的未来状态表示 world model world models world action model
71 Leveraging Human-In-The-Loop Demonstrations in Reinforcement Learning for Digital Twin-Driven Robot Flexibility 提出人机协作在线训练框架以提升机器人适应性 reinforcement learning imitation learning
72 LeWAM: End-to-End World and Action Modeling with JEPAs 提出LeWAM以解决控制任务中的世界与动作建模问题 behavior cloning flow matching world model

🔬 支柱三：空间感知与语义 (Perception & Semantics) (5 篇)

#题目一句话要点标签🔗⭐
73 2DGS-Planner: Rasterization-based Path Planning in 2D Gaussian Splatting Map 提出2DGS-Planner以解决2D高斯图中的路径规划问题 gaussian splatting splatting
74 STAG: A Sparse Traversability-Aware Graph Representation from Grid-Based Costmaps for Robotic Navigation 提出STAG以解决机器人导航中的高效全局规划问题 traversability
75 CoCam4D: Geometry-Aware Cooperative 4D Perception for Camera-Only Autonomous Driving 提出CoCam4D以解决摄像头感知的几何不确定性问题 depth estimation monocular depth VGGT
76 From Language to Motion: Task-Conditioned Focal-Stack Trajectory Integration for Microscopic Robots 提出语义到物理的框架以解决微型机器人任务几何问题 open-vocabulary open vocabulary
77 Multimodal Image-based Robotic Assistance and Guidance Environment (MIRAGE): A Novel Unified Framework for Surgical Navigation and Visualization 提出MIRAGE框架以解决外科手术导航与可视化问题 scene reconstruction multimodal

🔬 支柱九：具身大模型 (Embodied Foundation Models) (3 篇)

#题目一句话要点标签🔗⭐
78 ARC: A Reasoning Recipe for Robot Foundation Models 提出ARC以提升机器人基础模型的零-shot任务性能 VLA foundation model
79 Verifier-Mediated Multi-LLM Coordination for Multi-Robot Path Planning 提出验证者介导的多LLM协调方法以解决多机器人路径规划问题 large language model foundation model
80 What Should Embodied Agents Remember? Benchmarking and Evaluating Memory Modality for Embodied Agents 提出MeMoREval基准以优化具身智能体的记忆使用 large language model

🔬 支柱七：动作重定向 (Motion Retargeting) (1 篇)

#题目一句话要点标签🔗⭐
81 Distributed Relative Localization Based on Ultra-WideBand and LiDAR for Multi-robot with Limited Communication 提出基于UWB和LiDAR的分布式相对定位方法以解决多机器人定位问题 spatial relationship

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

#题目一句话要点标签🔗⭐
82 Fixed-Reference Pose Residuals for Measuring Cross-Dataset Cue Transfer in Human-Robot Interaction Anticipation 提出固定参考姿态残差模型以解决人机交互预测问题 egocentric ✅

⬅️ 返回 cs.RO 首页 · 🏠 返回主页