cs.RO（2026-09-21）

📊 共 56 篇论文 | 🔗 6 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (40 🔗4) 支柱九：具身大模型 (Embodied Foundation Models) (6) 支柱三：空间感知与语义 (Perception & Semantics) (5 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (5)

🔬 支柱一：机器人控制 (Robot Control) (40 篇)

#题目一句话要点标签🔗⭐
1 Opt2VLA: Force-Aware Vision-Language-Action for Contact-Rich Humanoid Whole-Body Manipulation 提出Opt2VLA以解决人形机器人接触丰富任务中的力控制问题 humanoid humanoid robot whole-body control
2 Imagine-RL: Residual-Confidence-Guided Cross-Attention for World-Model-Augmented VLA Reinforcement Learning 提出Imagine-RL以解决接触丰富操作中的可靠动作评估问题 manipulation reinforcement learning world model
3 Think Like a World Model, Act Like a VLA: Distilling World-Model Representations into Compact Robot Policies 提出通过世界模型蒸馏提升机器人政策的有效性 humanoid manipulation bi-manual
4 Smoothness as a Constraint for Stable Humanoid Locomotion 提出DeCap以解决人形机器人平稳行走问题 humanoid humanoid robot humanoid locomotion
5 PredActor: Predictive Action Diffusion for Steerable Onboard Humanoid Control 提出PredActor以解决人形机器人控制中的反馈响应问题 humanoid humanoid control Unitree
6 vla.simd: Efficient CPU Inference for Language-Conditioned Manipulation 提出vla.simd以解决无GPU环境下语言条件操作的高效推理问题 manipulation VLA language conditioned
7 Dexterous Robot Manipulation from Human Demonstrations via Contact-Anchored Retargeting and Residual Policy Learning 提出一种新方法以解决人类示范数据不足的问题 manipulation dexterous manipulation bi-manual
8 Bridge3D: Enabling Vision-Language-Action Models to See and Act in 3D 提出Bridge3D以解决VLA模型在3D操作中的局限性 manipulation vision-language-action VLA
9 LIBERO-VPro: Benchmarking Closed-Loop Visual Robustness of Robotic Foundation Models 提出LIBERO-VPro以评估机器人基础模型的视觉鲁棒性 manipulation vision-language-action VLA
10 Phrase-Level Robotic Guqin Performance: Bimanual Motion Planning and Audio-Tactile Interaction Monitoring 提出双臂机器人系统以实现自主的古琴演奏 humanoid humanoid robot manipulation
11 ARSTAG: An Agentic Real2Sim2Real System for Task-Specific Robot Data Generation 提出ARSTAG以解决机器人任务特定数据生成问题 manipulation dual-arm sim-to-real ✅
12 MIGU: Multimodal Instruction Grounding under Uncertainty for Manipulation Planning 提出MIGU框架以解决多模态指令理解中的不确定性问题 manipulation mobile manipulation motion planning
13 Automatic Labelling for Bimanual Mobile Manipulation 提出自动标注管道以解决双手移动操控任务的语义标注问题 manipulation mobile manipulation bi-manual
14 DexTacWAM: A Visuo-Tactile World-Action Model for Dexterous Manipulation 提出DexTacWAM以解决视觉与触觉信息融合不足的问题 manipulation dexterous manipulation bi-manual
15 MimicAgent: Quadruped Skills via Prompt-to-Trajectory Generation 提出MimicAgent以解决四足机器人技能学习问题 quadruped humanoid locomotion
16 CARE: Experience-Guided Atomic Corrective Execution for Vision-Language-Action Policies 提出CARE框架以解决视觉-语言-动作策略执行中的恢复问题 manipulation dual-arm vision-language-action ✅
17 D-JEPA: A Decision-Aligned Latent World Model 提出D-JEPA以解决决策对齐的潜在世界模型问题 manipulation world model world models
18 LLM-based Conversational AI Knowledge Assistant for MyBuddy Humanoid Robot 提出基于LLM的对话AI知识助手以提升人形机器人交互能力 humanoid humanoid robot large language model
19 Zeva-Ego: Egocentric Mid-Training with In-Context Causal Learning for Robot Manipulation 提出Zeva-Ego以解决机器人操控中的知识转化与适应性问题 manipulation egocentric VLA
20 Tactile-JEPA: Topology-Aware Self-Supervised Representation Learning for Distributed Tactile Sensors 提出Tactile-JEPA以解决分布式触觉传感器的自监督学习问题 manipulation dexterous manipulation policy learning ✅
21 Visuomotor Robotic Pruning in Planar Orchards Using Hybrid Reinforcement Learning 提出混合强化学习方法以实现平面果园的机器人修剪 sim-to-real motion planning reinforcement learning
22 TACIT: Tactile Contact Supervision for Spatial Attention in Dexterous Manipulation 提出TACIT以解决视觉运动策略在灵巧操作中的空间注意力问题 manipulation dexterous manipulation diffusion policy
23 DualWAM: Dual-System World Action Models for Asynchronous Global Planning and Local Refinement 提出DualWAM以解决机器人控制中的响应性问题 manipulation world action model world action models
24 H2RBench: A Real-to-Sim Benchmark for Evaluating Human-to-Robot Transfer 提出H2RBench以解决人机转移评估标准化问题 manipulation real2sim human-to-robot
25 RoboTalk: Learning Multi-Robot Communication and Coordination from Multimodal Demonstrations 提出RoboTalk以解决多机器人协作中的通信与协调问题 manipulation mobile manipulation multimodal
26 ActiveArena: Benchmarking and Understanding Active Perception in Robotic Manipulation 提出ActiveArena以解决机器人主动感知与操控评估问题 manipulation vision-language-action VLA
27 MIRA: Real-Time Full-Duplex Human-Robot Interaction for Embodied Companions 提出MIRA框架以解决实时人机交互中的响应同步问题 humanoid humanoid robot motion generation
28 AquaOrbit: Sim-to-Real Reinforcement Learning for Underwater Target Orbiting under Intermittent Visual Feedback 提出AquaOrbit以解决水下目标轨道控制中的视觉反馈中断问题 sim-to-real reinforcement learning
29 Learning Beyond What Humans Can Demonstrate 提出GLIDE框架以解决人类难以示范的机器人操作问题 manipulation teleoperation VR teleoperation
30 Object-Centric Conditioning for Visuomotor Flow Matching 提出SlotFlow以解决机器人视觉运动策略的鲁棒性问题 manipulation flow matching
31 From Semantic Decisions to Feasible Trajectories: Self-Evolving LLM-Guided Optimal Control for Narrow-Space Parking 提出SE-LLM-OCP以解决狭窄空间自动停车问题 trajectory optimization large language model
32 Safety Control of a Hyper-redundant Robot via Adaptive Weighted Control Barrier Functions 提出加权控制障碍函数以解决超冗余机器人安全控制问题 manipulation teleoperation
33 Touch2Robot: Robot Touch in the Human Demonstration Loop 提出Touch2Robot以解决机器人触觉反馈不足的问题 manipulation diffusion policy
34 InsertAnything: Generalizable Contact-Rich Precision Insertion from Simulation to Reality 提出一种强化学习框架以解决接触丰富的精确插入问题 manipulation reinforcement learning ✅
35 GraspTune: Tactile-Driven Execution Refinement for Robust Grasping 提出GraspTune以解决稳定抓取执行阶段挑战 manipulation PPO
36 Anticipatory Robot Goalkeeping via Monotone Optimal Stopping 提出单调最优停止方法以解决机器人守门员的预判问题 quadruped reinforcement learning
37 CAST: Collision-Aware Assembly with Construction Robots using Simultaneous Trajectory Estimation and Planning 提出CAST框架以解决建筑机器人碰撞感知组装问题 motion planning
38 Learning tactile perception from high-bandwidth single-point sensing 提出SpectRobot框架以解决单点触觉感知问题 manipulation
39 ScaleMPA: Rethinking Scalable RRT* Acceleration With a Grid-Native Representation 提出ScaleMPA以解决高维环境中的实时运动规划问题 motion planning
40 Robotic Valve Turning: Axial Misalignment Correction Using Reaction Torque Feedback 提出基于反作用力反馈的控制方法以纠正机器人阀门操作中的轴向失调问题 manipulation

🔬 支柱九：具身大模型 (Embodied Foundation Models) (6 篇)

#题目一句话要点标签🔗⭐
41 StenoVLA-3D: 3D-Aware Reasoning VLA for Navigation Through Gastrointestinal Stenoses 提出StenoVLA-3D以解决内窥镜导航中的狭窄区域问题 vision-language-action VLA
42 Uranus: Building the Next-Generation Simulation Infrastructure for Embodied AI 提出Uranus以解决机器人仿真基础设施不足问题 embodied AI
43 FoldQuantVLA: Native Low-Bit Quantization of Vision-Language-Action Models via Consistent Folding 提出FoldQuantVLA以解决低比特视觉-语言-动作模型的延迟问题 vision-language-action
44 What do VLM-Based Vision-Language Navigation Models Rely on: Interpreting and Steering Policy Behavior 提出可解释的视觉语言导航模型以提升导航决策透明度 VLN multimodal
45 ME-Brain-1.0: Memory, Cognition and Action for Evolving Embodied Intelligence 提出ME-Brain以解决现有智能体学习能力固定的问题 multimodal
46 Toward Human-in-the-Loop Robot Failure Recovery: Bridging Communication Gaps in Human-Robot Collaboration 提出LD-HRI以解决人机协作中的沟通差异问题 large language model

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

#题目一句话要点标签🔗⭐
47 Learning to Drive on Mars: Visual Multimodal Traversability Estimation for Off-World Navigation 提出一种不确定性感知的可通行性估计框架以解决火星自主导航问题 traversability multimodal ✅
48 A Topological Representation with Object-Path Graphs for Open-Vocabulary Instance Navigation 提出对象-路径图以解决开放词汇实例导航问题 open-vocabulary open vocabulary
49 OpenFlyScan: A Quality-Guided Aerial Reconstruction System for Consumer Drones 提出OpenFlyScan以解决消费级无人机的城市资产重建问题 3D gaussian splatting 3DGS gaussian splatting ✅
50 SE(3) Neural Potential Fields for 6-DoF Trajectory Planning Directly from Images Without Explicit 3D Reconstruction 提出SE(3)神经势场以解决无显式3D重建的6自由度轨迹规划问题 3D reconstruction
51 BayesianGS-SLAM: Uncertainty-Aware Neural Rendering SLAM via Probabilistic Formulation 提出BayesianGS-SLAM以解决神经渲染SLAM中的不确定性问题 3D gaussian splatting gaussian splatting splatting

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

#题目一句话要点标签🔗⭐
52 What Matters in Designing World Action Models: An Empirical Study 通过系统性研究设计世界行动模型以提升机器人控制能力 world model world models world action model
53 Beyond Visual Quality: A Study of Test-Time Planning with World Action Models 提出基于世界行动模型的测试时规划方法以提升决策质量 world action model world action models
54 Audio-based UAV Localization with Adaptive Temporal Correspondence via Reinforcement Learning 提出自适应时间对应的音频无人机定位方法以解决时效性问题 reinforcement learning Mamba
55 Performance-Preserving Online Adaptation in Social Navigation via Diffusion Steering 提出扩散引导的在线适应方法以解决社交导航中的性能保持问题 reinforcement learning deep reinforcement learning diffusion policy
56 Steerable and Reactive Grasping Through Modular Design with a Three-Point Interface 提出模块化设计的三点接口以实现灵活和反应式抓取 reinforcement learning affordance

⬅️ 返回 cs.RO 首页 · 🏠 返回主页