cs.RO（2026-10-02）

📊 共 36 篇论文 | 🔗 4 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (30 🔗4) 支柱二：RL算法与架构 (RL & Architecture) (3) 支柱三：空间感知与语义 (Perception & Semantics) (2) 支柱九：具身大模型 (Embodied Foundation Models) (1)

🔬 支柱一：机器人控制 (Robot Control) (30 篇)

#题目一句话要点标签🔗⭐
1 FastOPD: On-Policy Distillation for Lightweight VLA Deployment 提出FastOPD以解决VLA模型部署中的计算成本问题 manipulation world action model world action models
2 PointWAM: 3D World Action Modeling for Dexterous Robotic Manipulation 提出PointWAM以解决3D空间中灵巧机器人操作问题 manipulation dexterous manipulation world action model
3 DexJoCo-X: Benchmarking Action Representations for Multi-Hand Dexterous Manipulation 提出DexJoCo-X以解决多手灵巧操控的统一表示问题 manipulation dexterous hand dexterous manipulation
4 MobiAgent: Dual-Loop Recursive Policy Self-Improvement for Long-Horizon Mobile Manipulation 提出MobiAgent以解决长时间移动操控中的执行错误问题 locomotion manipulation mobile manipulation
5 ManiPhysicsBench: Physics-Based Assessment of Object Preservation in VLA Manipulation 提出ManiPhysicsBench以评估VLA操作中的物体保护问题 manipulation vision-language-action VLA
6 SimpleTouch: Can Vision-Language-Action Models Master Contact-Rich Manipulation Without Tactile Policy Pretraining? 提出SimpleTouch以简化触觉策略预训练在接触丰富操控中的应用 manipulation vision-language-action VLA
7 KungfuAthleteBot: learning high-dynamic humanoid motion from video with unified robust recovery 提出KungfuAthleteBot以解决高动态人形机器人运动学习问题 humanoid humanoid robot fall recovery
8 MixVLA: Adaptive Mixing of Non-Invariant Information for Generalizable Vision-Language-Action Models 提出MixVLA以解决VLA模型在OOD条件下的泛化问题 manipulation vision-language-action VLA
9 Around the World: Unified Learned Locomotion on a 270 g Continuous-Rotation Quadruped 提出统一学习的运动控制方法以解决亚千克四足机器人运动问题 quadruped locomotion reinforcement learning ✅
10 XGenAct: Geometry-Enhanced World Action Models through Cross-Task Generation 提出XGenAct以解决机器人操作中的空间理解问题 manipulation world action model world action models
11 EyeRobot 2.0: Active Gaze for Precise Manipulation without Wrist Cameras 提出EyeRobot 2.0以解决单摄像头精细操作问题 manipulation bi-manual bimanual manipulation
12 Equivariant Visual-Tactile Diffusion Policy for Contact-Rich Manipulation 提出VISTA以解决接触丰富操控中的样本效率问题 manipulation imitation learning diffusion policy
13 SARI: Phase-Split Sim-Real Co-Training for Contact-Rich Manipulation 提出SARI框架以解决接触丰富操作中的数据收集问题 manipulation vision-language-action VLA
14 Skill2Real: Agentic Skill Learning for Zero-Shot Sim-to-Real Robot Manipulation 提出Skill2Real框架以解决机器人技能从模拟到现实的转移问题 manipulation sim-to-real
15 DR-IPC: Disturbance-Resilient Integrated Planning and Control for LiDAR-Based Quadrotor Navigation 提出DR-IPC以解决LiDAR四旋翼在干扰下的导航问题 model predictive control trajectory optimization motion generation ✅
16 Beyond Reward Hacking: Proxy Divergence Across Four Layers of a Staged Humanoid Learning Pipeline 提出四层代理收敛机制以解决奖励黑客问题 legged robot humanoid reinforcement learning
17 RoboBridge: A Self-Evolving Embodied Agent Framework for Sim-to-Real Transfer 提出RoboBridge框架以解决仿真到现实转移的适应性问题 manipulation sim-to-real vision-language-action
18 Learning Reflexive Behavior for Contact-Rich Manipulation 提出反射行为学习以解决接触丰富的操作问题 manipulation dual-arm penetration
19 AdaTempo: Learning Shared Relative Tempo from Demonstrations for Faster Robot Manipulation 提出AdaTempo以解决机器人操作中的速度加速问题 manipulation imitation learning diffusion policy
20 LLA-MPC on Embedded Hardware: Rapid Adaptive Control with Thousands of Parallel Models 提出LLA-MPC以解决实时自适应控制问题 MPC model predictive control
21 RATE: Risk-Aware Tactile Encoding for Contact-rich Robotic Manipulation 提出风险感知触觉编码RATE以解决机器人操作中的模糊性问题 manipulation representation learning
22 DeltaWorld: Physically Consistent Interactive World Simulators via Action-Conditioned Latent Increment Learning 提出DeltaWorld以解决机器人操控中的物理一致性问题 manipulation world model world models
23 World Action Learning via Interaction-Centric Spectral Latent Guidance 提出WING框架以解决机器人政策学习中的数据收集难题 manipulation egocentric cross-embodiment ✅
24 A Unified Framework for Empowerment and Predictive Control 提出统一框架以结合赋权与预测控制 MPC model predictive control trajectory optimization
25 RoboChemGym: A Protocol-Driven Generative Simulation Framework for Long-Horizon Chemical Manipulation 提出RoboChemGym以解决化学实验自动化数据不足问题 manipulation
26 Bridging Frontier Reasoning and Robot Execution: From Autonomous Demonstration Generation to Dense Language Supervision 提出自主演示生成与密集语言监督以解决机器人实时控制问题 manipulation instruction following
27 LOCUS: Landmark-Oriented Container Discrimination Using Spatial Graphs 提出LOCUS以解决复杂空间语义关系下的容器识别问题 manipulation physically plausible
28 Self-Repairing Recurrent Ensembles for Real-Time Recovery from Distribution Shift 提出自修复递归集成方法以应对分布漂移问题 teleoperation imitation learning
29 Safe Streaming Flow Planning by Aligning Sampling Dynamics with Execution Dynamics 提出SafeStreamingFlow以解决安全流规划问题 locomotion flow matching
30 Proprioceptive Sketches as Long-Horizon Intent for Generative Action Policies 提出Proprioceptive Action Models以解决长时间意图预测问题 bi-manual ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (3 篇)

#题目一句话要点标签🔗⭐
31 AVL-JEPA: Preventing Causal Dynamics Information Collapse In Joint Embedding Predictive Architecture World Models 提出AVL-JEPA以解决因果动态信息崩溃问题 world model world models JEPA
32 From Language Priors to Field Adaptation: Preference Learning for Traversability Estimation 提出样本高效的偏好学习以解决可通行性估计问题 preference learning semantic map traversability
33 Permutation Robustness Is Not Enough: Action Collapse in Multi-Agent Transformer Policies 提出多代理变换器政策评估方法以解决行动崩溃问题 PPO transformer policy

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

#题目一句话要点标签🔗⭐
34 GeoScaffold: Learning Compact Geometric Latents via Reconstruction for Efficient Vision-Language Navigation 提出GeoScaffold框架以解决视觉语言导航中的几何信息不足问题 traversability egocentric VLN
35 CORNAV: Construction-Aware Reasoning for Robot Navigation on Active Worksites 提出CORNAV以解决建筑工地机器人导航问题 scene understanding open-vocabulary open vocabulary

🔬 支柱九：具身大模型 (Embodied Foundation Models) (1 篇)

#题目一句话要点标签🔗⭐
36 Detect and Suppress: A Mechanistic Defense against Adversarial Patches in VLA Models 提出机制性防御以应对VLA模型中的对抗性补丁问题 vision-language-action VLA

⬅️ 返回 cs.RO 首页 · 🏠 返回主页