cs.RO（2026-05-19）

📊 共 27 篇论文 | 🔗 9 篇有代码

🎯 兴趣领域导航

支柱一：机器人控制 (Robot Control) (14 🔗5) 支柱二：RL算法与架构 (RL & Architecture) (5) 支柱三：空间感知与语义 (Perception & Semantics) (4) 支柱九：具身大模型 (Embodied Foundation Models) (4 🔗4)

🔬 支柱一：机器人控制 (Robot Control) (14 篇)

#题目一句话要点标签🔗⭐
1 CEER: Compliant End-Effector and Root Control as a Unified Interface for Hierarchical Humanoid Loco-Manipulation 提出CEER，统一末端执行器与根控制接口，实现人型机器人分层操作控制。 humanoid humanoid robot whole-body control
2 PAPO-VLA: Planning-Aware Policy Optimization for Vision-Language-Action Models 提出PAPO-VLA，提升视觉-语言-动作模型在机器人操作任务中的可靠性 manipulation vision-language-action VLA
3 DEFLECT: Delay-Robust Execution via Flow-matching Likelihood-Estimated Counterfactual Tuning for VLA Policies DEFLECT：通过流匹配似然估计反事实调整，实现视觉-语言-动作策略的延迟鲁棒执行 bi-manual flow matching vision-language-action
4 RoVLA: Multi-Consistency Constraints for Robust Vision-Language-Action Models 提出RoVLA以解决视觉语言行动模型的脆弱性问题 manipulation predictive model vision-language-action ✅
5 Beyond Binary Success: A Diagnostic Meta-Evaluation Framework for Fine-Grained Manipulation 提出MetaFine诊断框架，解决细粒度操作中二元评价指标掩盖模型瓶颈的问题。 manipulation embodied AI vision-language-action ✅
6 TravExplorer: Cross-Floor Embodied Exploration via Traversability-Aware 3-D Planning TravExplorer：基于可通行性三维规划的跨楼层具身探索 trajectory optimization Unitree open-vocabulary ✅
7 ARC-RL: A Reinforcement Learning Playground Inspired by ARC Raiders 提出ARC-RL：一个受ARC Raiders启发的强化学习腿部运动控制环境 quadruped legged locomotion locomotion
8 RoHIL: Robust Human-in-the-Loop Robotic Reinforcement Learning Against Illumination Variations RoHIL：一种针对光照变化的鲁棒人机协作强化学习方法 manipulation reinforcement learning world model ✅
9 Beyond Action Residuals: Real-World Robot Policy Steering via Bottleneck Latent Reinforcement Learning 提出ZPRL，通过瓶颈潜在空间强化学习引导预训练机器人策略，提升真实世界操作性能。 manipulation reinforcement learning flow matching ✅
10 Graph Neural Planning and Predictive Control for Multi-Robot Communication-Constrained Unlabeled Motion Planning 提出基于图神经网络规划与预测控制的多机器人通信受限运动规划方法 model predictive control motion planning
11 KIO-planner: Attention-Guided Single-Stage Motion Planning with Dual Mapping for UAV Navigation KIO-planner：基于注意力机制和双重映射的无人机单阶段运动规划 motion planning
12 Bilateral Teleoperation with Compliant 6-DOF Pose-and-Force Sensing 提出基于低成本合规6-DOF传感器的双向遥操作框架以解决现有平台的局限性 teleoperation
13 Self-assembling Modular Aerial Robot for Versatile Aerial Tasks 提出LEGION自组装模块化飞行机器人，实现灵活导航与鲁棒空中操作的结合 manipulation
14 ContextFlow: Hierarchical Task-State Alignment for Long-Horizon Embodied Agents ContextFlow：用于长时程具身智能体任务状态对齐的分层框架 manipulation

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

#题目一句话要点标签🔗⭐
15 FlyMirage: A Fully Automated Generation Pipeline for Diverse and Scalable UAV Flight Data via Generative World Model FlyMirage：基于生成世界模型全自动生成多样且可扩展的无人机飞行数据 world model world models 3D gaussian splatting
16 SafeAlign-VLA: A Negative-Enhanced Safe Alignment Framework for Risk-Aware Autonomous Driving SafeAlign-VLA：面向风险感知的自动驾驶负样本增强安全对齐框架 reinforcement learning vision-language-action VLA
17 HEAT: Heterogeneous End-to-End Autonomous Driving via Trajectory-Guided World Models HEAT：基于轨迹引导世界模型的异构端到端自动驾驶 world model world models
18 Beyond Imitation: Learning Safe End-to-End Autonomous Driving from Hard Negatives BeyondDrive：基于难负样本学习的安全端到端自动驾驶 imitation learning flow matching zero-shot transfer
19 Aerial Inspection Behaviors via RL-based Quadrotor Control for Under-canopy Forest Environments 提出基于强化学习的四旋翼控制方法，用于林下环境的自主巡检任务 reinforcement learning deep reinforcement learning

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

#题目一句话要点标签🔗⭐
20 Beyond Waypoints: Dual-Heatmap Grounding for Cross-Embodiment Semantic Navigation 提出基于双热图的跨具身语义导航框架，提升机器人环境交互能力 affordance cross-embodiment foundation model
21 CLUE: Adaptively Prioritized Contextual Cues by Leveraging a Unified Semantic Map for Effective Zero-Shot Object-Goal Navigation CLUE：利用统一语义地图自适应地调整上下文线索优先级，实现有效的零样本目标导航 semantic map large language model
22 Minimalist Visual Inertial Odometry 提出一种极简视觉惯性里程计，仅用四个光电二极管实现差速驱动机器人稳健的平面里程计。 VIO motion estimation
23 MCNav: Memory-Aware Dynamic Cognitive Map for Zero-shot Goal-oriented Navigation MCNav：面向零样本目标导航的记忆增强动态认知地图 scene understanding large language model

🔬 支柱九：具身大模型 (Embodied Foundation Models) (4 篇)

#题目一句话要点标签🔗⭐
24 D-CLING: Prior-Preserving Depth-Conditioned Fine-Tuning for Navigation Foundation Models D-CLING：深度条件微调导航基础模型，保持先验知识并提升泛化性 foundation model ✅
25 CANINE: Coaching Visually Impaired Users for Interactive Navigation with a Robot Guide Dog CANINE：通过个性化指导提升视障用户与导盲犬机器人的交互导航能力 foundation model ✅
26 RoboJailBench: Benchmarking Adversarial Attacks and Defenses in Embodied Robotic Agents RoboJailBench：用于具身机器人代理对抗攻击与防御的基准测试平台 embodied AI ✅
27 PRISM-SLAM: Probabilistic Ray-Grounded Inference for Scale-aware Metric SLAM 提出PRISM-SLAM以解决单目SLAM中的尺度模糊问题 foundation model ✅

⬅️ 返回 cs.RO 首页 · 🏠 返回主页