cs.RO（2025-12-10）

📊 共 23 篇论文 | 🔗 6 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (16 🔗4) 支柱三：空间感知与语义 (Perception & Semantics) (4 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (2) 支柱九：具身大模型 (Embodied Foundation Models) (1 🔗1)

🔬 支柱一：机器人控制 (Robot Control) (16 篇)

#	题目	一句话要点	标签	🔗	⭐
1	HiF-VLA: Hindsight, Insight and Foresight through Motion Representation for Vision-Language-Action Models	HiF-VLA：利用运动表征实现视觉-语言-动作模型中的回溯、洞察与前瞻能力	manipulation motion representation vision-language-action
2	One-Shot Real-World Demonstration Synthesis for Scalable Bimanual Manipulation	BiDemoSyn：基于单样本真实演示合成可扩展的双臂操作数据	manipulation bi-manual dual-arm
3	GLaD: Geometric Latent Distillation for Vision-Language-Action Models	GLaD：几何潜在蒸馏增强视觉-语言-动作模型的空间推理能力	manipulation distillation VGGT
4	Safe Learning for Contact-Rich Robot Tasks: A Survey from Classical Learning-Based Methods to Safe Foundation Models	综述：面向接触密集型机器人任务的安全学习方法，从经典方法到安全具身智能模型	manipulation reinforcement learning vision-language-action	✅
5	ReMoSPLAT: Reactive Mobile Manipulation Control on a Gaussian Splat	ReMoSPLAT：基于高斯溅射的移动操作机器人反应式控制	manipulation mobile manipulation ReMoS
6	Scene-agnostic Hierarchical Bimanual Task Planning via Visual Affordance Reasoning	提出基于视觉可供性的场景无关分层双臂任务规划框架	manipulation bi-manual bimanual manipulation
7	Simultaneous Tactile-Visual Perception for Learning Multimodal Robot Manipulation	提出TacThru-UMI，结合触觉视觉同步感知与Transformer扩散策略，提升机器人操作精度。	manipulation imitation learning diffusion policy
8	Push Smarter, Not Harder: Hierarchical RL-Diffusion Policy for Efficient Nonprehensile Manipulation	提出HeRD：一种层级RL-扩散策略，用于高效的非抓取操作	manipulation reinforcement learning diffusion policy	✅
9	REASAN: Learning Reactive Safe Navigation for Legged Robots	REASAN：面向复杂动态环境，学习腿式机器人反应式安全导航	legged robot locomotion reinforcement learning	✅
10	Openpi Comet: Competition Solution For 2025 BEHAVIOR Challenge	OpenPI Comet在BEHAVIOR挑战赛中获得亚军，通过系统性研究训练技巧和数据显著提升性能。	manipulation mobile manipulation embodied AI	✅
11	A Hierarchical, Model-Based System for High-Performance Humanoid Soccer	提出一种分层模型系统，助力人形机器人ARTEMIS在RoboCup 2024中获胜	humanoid humanoid robot locomotion
12	Fast Functionally Redundant Inverse Kinematics for Robotic Toolpath Optimisation in Manufacturing Tasks	提出快速功能冗余逆运动学算法，优化制造任务中机器人工具路径	manipulation
13	Py-DiSMech: A Scalable and Efficient Framework for Discrete Differential Geometry-Based Modeling and Control of Soft Robots	Py-DiSMech：基于离散微分几何的软机器人建模与控制高效框架	sim-to-real
14	Bridging the Basilisk Astrodynamics Framework with ROS 2 for Modular Spacecraft Simulation and Hardware Integration	提出Basilisk与ROS 2的轻量级桥接方案，用于模块化航天器仿真与硬件集成	model predictive control
15	ViTA-Seg: Vision Transformer for Amodal Segmentation in Robotics	ViTA-Seg：用于机器人非完整性分割的视觉Transformer	manipulation
16	Development of a Compliant Gripper for Safe Robot-Assisted Trouser Dressing-Undressing	针对老年人穿脱裤子辅助，提出一种安全顺应性机器人夹持器	manipulation

🔬 支柱三：空间感知与语义 (Perception & Semantics) (4 篇)

#	题目	一句话要点	标签	🔗	⭐
17	D$^2$GSLAM: 4D Dynamic Gaussian Splatting SLAM	提出D$^2$GSLAM以解决动态环境下SLAM问题	gaussian splatting splatting scene reconstruction
18	YOPO-Nav: Visual Navigation using 3DGS Graphs from One-Pass Videos	YOPO-Nav：利用单次视频的3DGS图进行视觉导航	3D gaussian splatting 3DGS gaussian splatting
19	LISN: Language-Instructed Social Navigation with VLM-based Controller Modulating	提出LISN-Bench与Social-Nav-Modulator，实现语言指导的社交导航。	scene understanding instruction following	✅
20	Inertial Magnetic SLAM Systems Using Low-Cost Sensors	提出基于低成本惯性磁传感器的惯性磁SLAM系统，解决弱光环境定位问题。	visual odometry

🔬 支柱二：RL算法与架构 (RL & Architecture) (2 篇)

#	题目	一句话要点	标签	🔗	⭐
21	COVLM-RL: Critical Object-Oriented Reasoning for Autonomous Driving Using VLM-Guided Reinforcement Learning	提出COVLM-RL框架，利用VLM引导强化学习解决自动驾驶泛化性问题	reinforcement learning chain-of-thought
22	Generalizable Collaborative Search-and-Capture in Cluttered Environments via Path-Guided MAPPO and Directional Frontier Allocation	提出PGF-MAPPO，解决复杂环境下协作搜索捕获任务中探索效率低和泛化性差的问题。	reinforcement learning reward shaping

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

#	题目	一句话要点	标签	🔗	⭐
23	Token Expand-Merge: Training-Free Token Compression for Vision-Language-Action Models	提出TEAM-VLA，一种免训练的token压缩框架，加速VLA模型推理并保持性能。	vision-language-action VLA multimodal	✅

⬅️ 返回 cs.RO 首页 · 🏠 返回主页