cs.RO（2025-09-24）

📊 共 30 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (19 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (6 🔗1) 支柱三：空间感知与语义 (Perception & Semantics) (3) 支柱九：具身大模型 (Embodied Foundation Models) (2)

🔬 支柱一：机器人控制 (Robot Control) (19 篇)

#题目一句话要点标签🔗⭐
1 VisualMimic: Visual Humanoid Loco-Manipulation via Motion Tracking and Generation VisualMimic：基于视觉的人形机器人运动跟踪与生成，实现零样本真实环境操作 humanoid humanoid robot whole-body control
2 DynaFlow: Dynamics-embedded Flow Matching for Physically Consistent Motion Generation from State-only Demonstrations DynaFlow：嵌入动力学的Flow Matching用于从状态演示中生成物理一致的运动 quadruped flow matching motion generation
3 MARG: MAstering Risky Gap Terrains for Legged Robots with Elevation Mapping MARG：基于高程地图的四足机器人崎岖地形（间隙）安全穿越 quadruped legged robot locomotion
4 Diffusion-Based Impedance Learning for Contact-Rich Manipulation Tasks 提出基于扩散模型的阻抗学习框架，用于接触式操作任务 parkour manipulation teleoperation
5 Beyond Human Demonstrations: Diffusion-Based Reinforcement Learning to Generate Data for VLA Training 提出基于扩散模型的强化学习算法，为VLA模型生成高质量训练数据。 manipulation reinforcement learning diffusion policy
6 Parse-Augment-Distill: Learning Generalizable Bimanual Visuomotor Policies from Single Human Video 提出PAD框架，从单个人类视频学习可泛化的双臂操作策略 bi-manual sim-to-real motion planning ✅
7 HL-IK: A Lightweight Implementation of Human-Like Inverse Kinematics in Humanoid Arms HL-IK：一种轻量级的人形机器人手臂类人逆运动学实现 humanoid humanoid robot teleoperation
8 Robot Trajectron V2: A Probabilistic Shared Control Framework for Navigation 提出Robot Trajectron V2，用于导航的概率共享控制框架 shared control reinforcement learning multimodal
9 Learning-Based Collaborative Control for Bi-Manual Tactile-Reactive Grasping 提出基于学习的触觉反馈双臂协作抓取方法，提升对易碎和形变物体的抓取能力。 bi-manual MPC model predictive control
10 Generalist Robot Manipulation beyond Action Labeled Data 提出一种利用无动作标签数据的通用机器人操作方法，提升泛化能力。 manipulation open-vocabulary open vocabulary
11 TopoCut: Learning Multi-Step Cutting with Spectral Rewards and Discrete Diffusion Policies TopoCut：提出基于谱奖励和离散扩散策略的多步切割学习框架。 manipulation policy learning diffusion policy
12 Action-Informed Estimation and Planning: Clearing Clutter on Staircases via Quadrupedal Pedipulation 提出交互感知的状态估计与规划方法，解决四足机器人楼梯杂物清理难题 quadruped manipulation
13 BBoE: Leveraging Bundle of Edges for Kinodynamic Bidirectional Motion Planning BBoE：利用边束的运动学双向运动规划，提升复杂环境下的规划效率 motion planning
14 C-3TO: Continuous 3D Trajectory Optimization on Neural Euclidean Signed Distance Fields C-3TO：基于神经欧几里德符号距离场的连续3D轨迹优化 trajectory optimization
15 mindmap: Spatial Memory in Deep Feature Maps for 3D Action Policies 提出mindmap，利用深度特征图中的空间记忆实现3D动作策略，提升机器人操作能力 manipulation diffusion policy
16 An effective control of large systems of active particles: An application to evacuation problem 结合强化学习与人工势场的领导者控制策略，解决大规模主动粒子系统的疏散问题 manipulation reinforcement learning ✅
17 SAGE:State-Aware Guided End-to-End Policy for Multi-Stage Sequential Tasks via Hidden Markov Decision Process SAGE：基于隐马尔可夫决策过程的状态感知引导端到端多阶段序列任务策略 manipulation imitation learning
18 A Biomimetic Vertebraic Soft Robotic Tail for High-Speed, High-Force Dynamic Maneuvering 提出一种仿生椎骨软体机器人尾部，用于高速、高力动态操作。 quadruped
19 Simultaneous estimation of contact position and tool shape with high-dimensional parameters using force measurements and particle filtering 提出基于力测量和粒子滤波的同时估计接触位置和高维工具形状参数的方法 manipulation

🔬 支柱二：RL算法与架构 (RL & Architecture) (6 篇)

#题目一句话要点标签🔗⭐
20 Discrete Diffusion for Reflective Vision-Language-Action Models in Autonomous Driving 提出ReflectDrive，通过离散扩散和反射机制实现自动驾驶安全轨迹生成。 reinforcement learning imitation learning vision-language-action
21 D3Grasp: Diverse and Deformable Dexterous Grasping for General Objects D3Grasp：面向通用物体的多样化和可变形灵巧抓取 reinforcement learning privileged information penetration
22 Latent Activation Editing: Inference-Time Refinement of Learned Policies for Safer Multirobot Navigation 提出基于隐空间激活编辑的框架，用于提升多无人机导航的安全性。 reinforcement learning world model large language model
23 RoboSSM: Scalable In-context Imitation Learning via State-Space Models RoboSSM：基于状态空间模型实现可扩展的上下文模仿学习 imitation learning SSM ✅
24 Selective Progress-Aware Querying for Human-in-the-Loop Reinforcement Learning 提出SPARQ，一种基于学习进度的选择性查询策略，提升人机协同强化学习效率。 reinforcement learning
25 LLM Trainer: Automated Robotic Data Generating via Demonstration Augmentation using LLMs LLM Trainer：利用LLM增强的自动化机器人数据生成，用于模仿学习。 imitation learning large language model

🔬 支柱三：空间感知与语义 (Perception & Semantics) (3 篇)

#题目一句话要点标签🔗⭐
26 Where Did I Leave My Glasses? Open-Vocabulary Semantic Exploration in Real-World Semi-Static Environments 提出一种开放词汇语义探索系统，用于真实半静态环境下的机器人导航与地图维护。 semantic map open-vocabulary open vocabulary
27 Boosting LiDAR-Based Localization with Semantic Insight: Camera Projection versus Direct LiDAR Segmentation 提出融合语义信息的LiDAR定位方法，提升复杂环境下移动系统的精度和鲁棒性。 Depth Anything multimodal
28 Queryable 3D Scene Representation: A Multi-Modal Framework for Semantic Reasoning and Robotic Task Planning 提出3D可查询场景表示，用于语义推理和机器人任务规划 scene understanding multimodal

🔬 支柱九：具身大模型 (Embodied Foundation Models) (2 篇)

#题目一句话要点标签🔗⭐
29 Boosting Zero-Shot VLN via Abstract Obstacle Map-Based Waypoint Prediction with TopoGraph-and-VisitInfo-Aware Prompting 提出基于抽象障碍地图的航点预测以提升零-shot视觉语言导航性能 VLN large language model foundation model
30 Large Language Models for 3D IC Space Planning 提出基于大语言模型的3D IC空间规划方法，优化芯片布局。 large language model

⬅️ 返回 cs.RO 首页 · 🏠 返回主页