Skip to content
具身智能数据

2026 年 9 月具身智能论文知识点索引

本索引从同级 spiders/arxiv/papers/ 的 28 个日期批次读取论文，并按每篇 published 字段筛选 2026-08-27 至 2026-09-27（UTC）。这段时间有 314 篇唯一论文；其中 278 篇与本站的数据采集、处理、对齐、标注、质量控制、评估或数据集使用相关，整理为 26 个独立主题。另外 36 篇保留在末尾的范围外清单，便于复核筛选口径。最近收录的发表时间为 2026-09-24；索引并不声称 9 月 25 日之后没有新论文。

314 篇论文经发表时间去重和数据相关性判断，归入 26 个知识点主题及范围外清单

归类与使用方法

Section titled “归类与使用方法”

每篇论文有一个主主题，避免同一篇文献在目录中重复出现；一篇论文涉及多个问题时，主题文档可交叉引用。主题页重点解释可复用的数据输入与输出、字段和标签、对齐方式、质量检查、评估边界，而不按论文篇幅逐篇复述方法。研究结论应回看所链接的论文原文，尤其要核对实验条件与适用场景。

爬虫的中文总结覆盖了部分论文，但不是所有总结都通过事实核验；因此主题文档以原始摘要或论文正文作为事实依据。爬虫抓取的论文配图没有逐张记录再使用许可，本网站使用原创数据流程示意图。

论文摘要和正文先映射到主题，再提炼可复用的数据契约、质检规则与评估问题

主题与论文来源

Section titled “主题与论文来源”

远程操作、示教采集与动作重定向

Section titled “远程操作、示教采集与动作重定向”

11 篇。阅读主题文档。

  • 2609.30249 · RAPID: Robot Agentic Programming from Demonstrations（2026-09-24）
  • 2609.26520 · MATE: Multi-Agent Virtual Teleoperation Platform for Humanoid Collaboration Data Collection（2026-09-22）
  • 2609.20659 · HIL-UMI: Bringing Human-in-the-Loop Post-Training of Vision-Language-Action Models to Universal Manipulation Interface（2026-09-17）
  • 2609.18763 · Gated Residual Body-Hand Coordination for Whole-Body Humanoid Teleoperation（2026-09-16）
  • 2609.16705 · The Robot Data Factory（2026-09-15）
  • 2609.16437 · XRoboToolKit-T: Teleoperation with High Stability and Precision with Tactile Sensing for Contact-rich Manipulation（2026-09-14）
  • 2609.15716 · Continuous Manifold-Decomposed Impedance Retargeting for Contact-Rich Imitation Learning（2026-09-14）
  • 2609.16041 · MR-GLi: Mixed Reality-Based Gripper-Linked Overlays for Underwater Robot Arm Teleoperation via Bilateral Control（2026-09-11）
  • 2609.16040 · Bi-MoDe: Bilateral Control-based Imitation Learning via Modifier-Conditioned Decoding for Modulation of Execution Speed and Contact Intensity（2026-09-11）
  • 2609.12384 · Understanding Whole-Body Robot Teleoperation Strategies Under Diverse Task Objectives and Constraints（2026-09-11）
  • 2609.06434 · Can People Distinguish Human and AI Agency in Humanoid Teleoperation? A Preliminary Study of Agency Perception（2026-09-06）

第一视角/多视角人体与人-物交互重建

Section titled “第一视角/多视角人体与人-物交互重建”

6 篇。阅读主题文档。

  • 2609.30187 · Ego-Exo4D Human Meshes Dataset: 4D Human Motion Reconstruction for Ego-Exo Captures（2026-09-24）
  • 2609.19119 · Track, Articulate, Act: Generating Articulation from Casual Human Videos（2026-09-16）
  • 2609.16684 · MEgoVista: Multi-view Ego-aware Motion Estimation for Metric 4D Hands and Head in the Wild（2026-09-15）
  • 2609.16518 · Beyond Gestures: Estimating Full Hand Pose and Contact Forces from Wrist-Worn Pressure Sensor Array（2026-09-15）
  • 2609.14615 · Open-UniMo: Towards Unified Motion-Language Understanding and Generation in the Open World（2026-09-13）
  • 2608.27407 · Reconstructing Humans and Objects in Interaction using Large Reconstruction Models（2026-08-27）

仿真、合成数据与真实迁移

Section titled “仿真、合成数据与真实迁移”

10 篇。阅读主题文档。

  • 2609.26795 · φ-RIE: From Photorealistic Reconstruction to Interactive Environments（2026-09-22）
  • 2609.24815 · Uranus: Building the Next-Generation Simulation Infrastructure for Embodied AI（2026-09-21）
  • 2609.21817 · A Sim-to-Real Integration Pipeline for Training and Deployment of Chunk-Based VLA Manipulation Policies（2026-09-18）
  • 2609.20582 · V2-STRep: VLM-Grounded Structured Task Representations for Reusable Robot Skills Acquired from Generated Videos（2026-09-17）
  • 2609.19137 · Dreaming the Sound of Contact: Leveraging Video and Audio Generation for Zero-Shot Force-Aware Manipulation and Data Generation（2026-09-16）
  • 2609.17521 · PhysStream: Streaming Physics-Grounded Video Generation with Structured Scene Memory and Fine-Grained Motion Control（2026-09-15）
  • 2609.16919 · NeuroSymbEAD: A Large Scale Neuro-Symbolic Caption Dataset for Omni-Directional Embodied Autonomous Driving（2026-09-15）
  • 2609.06279 · IM-ENGINE: Image Editing for Embodied Data Generation（2026-09-05）
  • 2609.02886 · SolarWM: Open Data and Scalable Training for Long-Horizon Video World Models（2026-09-02）
  • 2608.27073 · SpatialCrafter: Single Image World Modeling with Generative 3D Proxies（2026-08-27）

跨模态传感器与时空对齐

Section titled “跨模态传感器与时空对齐”

4 篇。阅读主题文档。

  • 2609.15509 · StereoPatch: Patch-Aligned RGB-Depth Fusion for Spatial Perception in Robot Manipulation（2026-09-14）
  • 2609.13053 · Dynin-Robotics: Omnimodal Unified Diffusion Vision-Language-Action Model（2026-09-11）
  • 2609.05361 · Development of a Humanoid Robot Prototype for Multimodal Human-Robot Interaction（2026-09-04）
  • 2608.30643 · Temporal Forcing: 4D Representation Alignment for Vision-Language-Action Models（2026-08-31）

三维几何、目标/部件定位与空间表征

Section titled “三维几何、目标/部件定位与空间表征”

11 篇。阅读主题文档。

  • 2609.28184 · VLMs Can Describe, But Not Measure: Object-Centric Scene Understanding for Robotic Manipulation（2026-09-23）
  • 2609.20586 · CoRef-GS: Cooperative Referring Gaussian Splatting for Multi-Agent Scene Understanding（2026-09-17）
  • 2609.19142 · PointZero: 3D Point Track Completion for Learning Transferable 3D Dynamics（2026-09-16）
  • 2609.13812 · GeomVLA: Unifying Scene, Motion, and Action in 3D（2026-09-12）
  • 2609.12898 · UniPart: Towards Zero-shot Language-Grounded 3D Part Segmentation for Embodied Interaction（2026-09-11）
  • 2609.12874 · VideoTok4D: A 4D-Aware Video Tokenizer for Compact World Representation（2026-09-11）
  • 2609.12721 · Improving Imitation Learning Efficiency for Manipulation through Geometric Prior Pretraining（2026-09-11）
  • 2609.06256 · GloVLA: Let Geometry Move and Local VLA Interact for Robust Object-Centric Manipulation in Unstructured Environments（2026-09-05）
  • 2609.04193 · GIFT: Guided Intermediate Feature Training via Action-Oriented Structural Supervision for Robotic Manipulation（2026-09-03）
  • 2608.30451 · SeqAlign3DVG: A Sequence-Aligned Benchmark and Voxel Reasoning Framework for 3D Visual Grounding（2026-08-31）
  • 2608.27226 · DINOcular: Self-Supervised Visuospatial Representations（2026-08-27）

主动感知、测量与采样

Section titled “主动感知、测量与采样”

5 篇。阅读主题文档。

  • 2609.21929 · MAAP: Multi-Agent Active Perception for Collaborative Manipulation（2026-09-18）
  • 2609.16786 · Optimal Excitation Trajectories for System Identification of Underwater Vehicles（2026-09-15）
  • 2609.14219 · Task-Specified Active Metrological Inspection with Measurement-Steered VLA Manipulation and Deterministic Evidence Gating（2026-09-13）
  • 2609.12894 · Before the Tipping Point: Force-Guided Active Perception for Shape-Agnostic Estimation of 3D Centers of Mass（2026-09-11）
  • 2608.27088 · Active sensing to characterize the heterogeneity of plant stress（2026-08-27）

潜在动作与跨本体数据对齐

Section titled “潜在动作与跨本体数据对齐”

8 篇。阅读主题文档。

  • 2609.21983 · SkelWAM: A Skeleton-Guided World-Action Model for Zero-Shot Cross-Embodiment Manipulation（2026-09-18）
  • 2609.21948 · GALA: Geometry-Aware Latent Action Modeling for Vision-Language-Action Model Pretraining across Embodiments（2026-09-18）
  • 2609.17099 · GeoLAM: Learning Geometry-Grounded Latent Actions from Unlabeled Human Videos（2026-09-15）
  • 2609.16815 · Rethinking Visual Embodiment Dependence in Visuomotor Policies（2026-09-15）
  • 2609.16641 · SAVLA: Symmetry-Aware Vision-Language-Action Models for Robotic Manipulation（2026-09-15）
  • 2609.16071 · Schema-Adaptive Action-Conditioned JEPA for Cross-Machine CNC Transfer under Partial Sensor Overlap（2026-09-13）
  • 2609.12641 · Breaking the Vision-Action Shortcut: Latent Interface Training for Generalizable Robotics Foundation Models（2026-09-11）
  • 2609.02634 · Latent Cluster Analysis for Vision-Language-Action Models（2026-09-02）

示教数据筛选、再利用与后训练

Section titled “示教数据筛选、再利用与后训练”

6 篇。阅读主题文档。

  • 2609.28314 · TANDEM: Task and Motion Planning with As-Needed Demonstrations for Efficient Vision-Language-Action Model Fine-tuning（2026-09-23）
  • 2609.26672 · Imperfection for Precision: Upcycling Imperfect Data for High-Precision Robotic Manipulation（2026-09-22）
  • 2609.24996 · Learning Beyond What Humans Can Demonstrate（2026-09-21）
  • 2609.19138 · In-Context Robot Learning with VLM Agents（2026-09-16）
  • 2609.13851 · ReWeight: Leveraging Human Data for VLA Post-Training via Demonstration Retrieval and Sample Weighting（2026-09-12）
  • 2609.12316 · DATAFARM: Distribution-Aligned Task and Motion Planning for Fine-Tuning Vision-Language-Action Models（2026-09-11）

模仿学习数据与动作生成

Section titled “模仿学习数据与动作生成”

4 篇。阅读主题文档。

  • 2609.30127 · Faster Visuomotor Policy Learning on Action Manifolds via Riemannian MeanFlow（2026-09-24）
  • 2609.15631 · Flow-Matched Motion Priors: Online Optimal-Transport Rewards for Imitation Learning（2026-09-14）
  • 2609.15162 · LieSpline-DP: Lie-Group B-Spline Diffusion Policy for Smooth Robot Manipulation（2026-09-14）
  • 2609.10405 · Frequency-Conditioned Flow Matching for Vision-Language-Action Models（2026-09-09）

世界模型训练数据与物理表征

Section titled “世界模型训练数据与物理表征”

22 篇。阅读主题文档。

  • 2609.30214 · Underwater C3-JEPA: An Object-Centric Cross-View World Model for ROV Salvage（2026-09-24）
  • 2609.28466 · The Past Frames the Future: Memory for Autoregressive Video Generation（2026-09-23）
  • 2609.28414 · Frozen Flows Forget: Diagnosing and Restoring Lost Motion in a Latent-flow World Model（2026-09-23）
  • 2609.28393 · PointCast: One World Model for Rigid, Articulated, and Deformable Object Manipulation（2026-09-23）
  • 2609.24984 · WorldCrafter: Consistent Video World Model with Implicit 3D-aware Memory（2026-09-21）
  • 2609.20709 · MoWAM: Explicit Future Motion Prediction for Efficient World Action Models（2026-09-17）
  • 2609.20669 · Learning Foresight without Explicit Trajectories for 3D Diffusion Policies（2026-09-17）
  • 2609.17372 · XPACE: Joint World and Action Modeling from Heterogeneous Experience（2026-09-15）
  • 2609.16697 · World Models for Embodied Intelligence: From Plausible to Controllable to Actionable（2026-09-15）
  • 2609.15570 · DIDO: Distilling Interaction-Centric Dynamics into One-Step Denoising for World Action Models（2026-09-14）
  • 2609.14973 · PhysBrain 1.5: From Vision-Language Models to Physical Foundation Models（2026-09-14）
  • 2609.16074 · World-Action Models for Robot Learning and Control: A Survey（2026-09-13）
  • 2609.14462 · AlayaVista: Streaming World Modeling from Panoramic States to Perspective Video（2026-09-13）
  • 2609.12441 · IMPLY: Physically Anchored Consistency for World-Model Rollouts（2026-09-11）
  • 2609.10540 · Programmable World Model（2026-09-09）
  • 2609.10506 · DUET-DINO: Simultaneous Cross-View World Modeling for Latent Planning in Robot Manipulation（2026-09-09）
  • 2609.10464 · Semigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics Generalization（2026-09-09）
  • 2609.06302 · CST-WM: A Causally Structured World Model for Embodied Visual Tracking（2026-09-05）
  • 2608.28491 · AcrossVAM1.0: Particle World Modeling for Text-Assisted Robot Video Prediction（2026-08-28）
  • 2608.27406 · CLAP: Cross-Embodiment Video World Models are Zero-Shot Physical Simulators（2026-08-27）
  • 2608.27367 · Successive Capacity Growth: Task-Complexity-Driven Width and Depth Expansion for Vision Transformer Encoders in JEPA World Models（2026-08-27）
  • 2608.27033 · Riemann-1.0: An Embodied World Action Model for Physical AI（2026-08-27）

世界模型规划与决策评估

Section titled “世界模型规划与决策评估”

12 篇。阅读主题文档。

  • 2609.30264 · AD-WM: Action-Discriminative World Models for Counterfactual Model Predictive Control（2026-09-24）
  • 2609.30247 · Rolling-WAM: World Action Models with Rolling Imagination（2026-09-24）
  • 2609.30036 · Aim Short to Reach Far: Your Frozen World Model Can Plan Better Than You Think（2026-09-24）
  • 2609.28431 · LiMA: Bridging Long-term Imagination to Real-time Dexterous Manipulation via Asynchronous Diffusion（2026-09-23）
  • 2609.28258 · Generalizable Robotic Insertion with World Models（2026-09-23）
  • 2609.24868 · DualWAM: Dual-System World Action Models for Asynchronous Global Planning and Local Refinement（2026-09-21）
  • 2609.24749 · D-JEPA: A Decision-Aligned Latent World Model（2026-09-21）
  • 2609.20575 · Accelerating Visual Policy Learning with Sampling-Based Model Predictive Control（2026-09-17）
  • 2609.15801 · When Should a World Model Move? Loss-Conditioned State Execution（2026-09-14）
  • 2609.15382 · From Prediction to Decision: World-Model-Guided Action Selection for Continuous Pile Excavation（2026-09-14）
  • 2609.14073 · LPA-CWM: A Learned Physical Adjudicator for Motion Reasoning with Counterfactual World Models（2026-09-12）
  • 2609.13845 · LePlanner: An Iterative Amortized Controller For World Models（2026-09-12）

长时记忆、经验检索与连续视频

Section titled “长时记忆、经验检索与连续视频”

9 篇。阅读主题文档。

  • 2609.28429 · Watch, Recall, Act: Always-On Robots in Concurrent Embodied Streams（2026-09-23）
  • 2609.28256 · MemBodied: Recurrent Associative Memory for Vision-Language-Action Models（2026-09-23）
  • 2609.22040 · PRIME: Perception Feedback with Situational Memory Embeddings in VLA Models（2026-09-18）
  • 2609.20820 · Workspace Models: Lightweight Robotic Memory via Saliency-Driven Supervision（2026-09-17）
  • 2609.16864 · TEMPO: Learning Temporal Context for Dynamic Robot Manipulation（2026-09-15）
  • 2609.15976 · MessyMem: Learning-from-Doing Memory for Mobile Manipulation（2026-09-14）
  • 2609.14561 · GLAM: Training a latent world model over global spatiotemporal memory for active exploration and navigation（2026-09-13）
  • 2609.02780 · ShallowStream: Index Shallow then Answer Deep for Streaming Video Understanding（2026-09-02）
  • 2608.27328 · R2M-Bench: Evaluating Revisit Memory via Relative Consistency in Interactive Video World Models（2026-08-27）

语言指令、视觉条件与动作对齐

Section titled “语言指令、视觉条件与动作对齐”

5 篇。阅读主题文档。

  • 2609.24995 · MIGU: Multimodal Instruction Grounding under Uncertainty for Manipulation Planning（2026-09-21）
  • 2609.15169 · GRAVA: Grounded Reasoning-to-Action Representation and Learning for Autonomous Driving（2026-09-14）
  • 2609.05376 · What Matters, When? Diagnosing and Improving Conditional Visual Grounding in Visuomotor Imitation Policies（2026-09-04）
  • 2609.05260 · One Word, Different Action: A Real-Robot Benchmark for Language-Conditioned Embodied Reasoning（2026-09-04）
  • 2609.02653 · HINT: Human-Intent Inception for Long-Horizon Robot Manipulation（2026-09-02）

任务分解、技能表征与组合

Section titled “任务分解、技能表征与组合”

11 篇。阅读主题文档。

  • 2609.26499 · Generalizing Manipulation Skills with a Local Coding Agent（2026-09-22）
  • 2609.18869 · KINO: A Keyframe Interface for VLM Planning and Whole-Body Control in Humanoid Loco-Manipulation（2026-09-16）
  • 2609.18813 · Asymptotically Optimal Multi-Robot Task and Motion Planning（2026-09-16）
  • 2609.17263 · CAD-Based Relation Learning and Geometric-Symbolic Planning for Robotic Assembly（2026-09-15）
  • 2609.16331 · ManiSkillFormer: Demonstration-Free Compositional Manipulation via Task-Conditioned Geometric Contracts（2026-09-14）
  • 2609.15587 · An Information-Space Perspective to Scene Graph Sufficiency for Robotic Task Planning（2026-09-14）
  • 2609.16056 · Managing Action Preconditions in Neuro-Symbolic RL: Three Placement Strategies for Embodied Agents（2026-09-13）
  • 2609.10522 · Show-Harness: Just a VLM Agent Can Play Robots（2026-09-09）
  • 2609.05369 · Towards Neuro-Symbolic Procedural Reasoning for Long-Horizon Vision-Language-Action Manipulation（2026-09-04）
  • 2608.31167 · SUN: Persistent Programs For Language-Grounded Control-to-Learning-to-Real Policies（2026-08-31）
  • 2608.27371 · Embodied Scene Rearrangement Planning（2026-08-27）

VLA 动作分块、推理延迟与实时执行

Section titled “VLA 动作分块、推理延迟与实时执行”

11 篇。阅读主题文档。

  • 2609.20776 · GeoAAC: Geometry-Based Adaptive Action Chunking from Denoising Trajectories in VLA Policies（2026-09-17）
  • 2609.20648 · SkipVLA: Skipping VLA Steps with Classical Planning for Fast Robot Manipulation（2026-09-17）
  • 2609.19104 · rMuscle: Robotic Muscle Memory for Efficient Vision-Language-Action Model Inference（2026-09-16）
  • 2609.17210 · FluxVLA Engine: A One-Stop VLA Engineering Platform for Embodied Intelligence（2026-09-15）
  • 2609.16503 · Dense to MoE Adaptation for Compact Vision Language Action Policies（2026-09-15）
  • 2609.15840 · Uncertainty-Guided Sparse Refinement for Action Chunking Transformer Policies（2026-09-14）
  • 2609.15322 · Planning in the Backbone: DiffAdapterVLA for Native Continuous Trajectory Generation with Driving VLMs（2026-09-14）
  • 2609.14146 · When Faster VLA Deployment Changes Closed-Loop Behavior: Task Success-Latency Analysis of SmolVLA Across PyTorch and ONNX Variants（2026-09-12）
  • 2609.13984 · What Makes an Efficient VLA? Navigating Action-Head Design, Scaling, and Latency（2026-09-12）
  • 2609.13695 · GROOVE: Geometry-Guided Reduction of Operational-Space Jerk in VLA Execution（2026-09-12）
  • 2608.27384 · FlashVLA: Streaming Action Decoding for Fast and Asynchronous VLA Inference（2026-08-27）

闭环反馈、失败检测与纠错

Section titled “闭环反馈、失败检测与纠错”

8 篇。阅读主题文档。

  • 2609.30092 · Self-Adaptive VLA for Robust Robot Deployment（2026-09-24）
  • 2609.30024 · Body-Grounded Replanning for Physically Adaptive Manipulation（2026-09-24）
  • 2609.21908 · CommitFlow: Semantic Commitment Verification and Local Correction for Long-Horizon Robot Manipulation VLA Execution（2026-09-18）
  • 2609.20646 · TraceFlow: Guiding Frozen Flow-Matching Robot Policies with Success and Failure Traces（2026-09-17）
  • 2609.17404 · Residual Fault Adaptation for Dexterous In-Hand Manipulation Under Runtime Joint Faults（2026-09-15）
  • 2609.14633 · REVOLVE: An Automated Closed-Loop Framework for Evolving Robot Manipulation with Minimal Human Intervention（2026-09-13）
  • 2609.06508 · VLA-Corrector: Stage-Aware Observable State Understanding for Prompt-Based Closed-Loop Recovery of Vision-Language-Action Policies（2026-09-06）
  • 2608.27079 · GRAFT: Grounded and Efficient Online Reinforcement Adaptation for Fine-Grained Robot Manipulation（2026-08-27）

触觉、力觉与接触数据

Section titled “触觉、力觉与接触数据”

16 篇。阅读主题文档。

  • 2609.30082 · Real-Time Force Regulation for Whole-Hand Dexterous Grasping（2026-09-24）
  • 2609.24976 · DexTacWAM: A Visuo-Tactile World-Action Model for Dexterous Manipulation（2026-09-21）
  • 2609.20761 · Agile-WAM: An Agile Tactile World Action Model for Contact-Rich Robot Control（2026-09-17）
  • 2609.20649 · DexTouch-WM: Learning Action-Conditioned Tactile World Models from Human Touch for Dexterous Robot Manipulation（2026-09-17）
  • 2609.16504 · UniDex-ViTac: Learning Unified Visuo-Tactile Dexterous Manipulation Policy from Human Video Data（2026-09-15）
  • 2609.15921 · Touch2Trace: Tactile-Driven Imitation Learning for Dexterous Cable Tracing（2026-09-14）
  • 2609.15910 · SlipSense: Multimodal Tactile Learning for Low-Latency and Generalized Slip Detection（2026-09-14）
  • 2609.15198 · PredTac: Learning Contact-Rich Manipulation with Predicted Touch（2026-09-14）
  • 2609.15012 · Atomic Motion Coordinate for Language-Steerable and Force-Responsive Manipulation（2026-09-14）
  • 2609.14156 · Visible Touch: Rendering Contact for Visuomotor Policies（2026-09-12）
  • 2609.14133 · Vision-Force Admittance Learning for Peg Insertion into a Movable Hole（2026-09-12）
  • 2609.13779 · Force-Aware Reinforcement Learning with Hybrid Sensorless Force Estimation for Wheeled-Legged Loco-Manipulation（2026-09-12）
  • 2609.12737 · Control Architecture for Safe Grasping of Fragile Objects Using a Coarse Position-Controlled Gripper（2026-09-11）
  • 2609.12549 · STAR: Sparse Tactile Representation Learning in Vision-Tactile-Language-Action Models for Dexterous Manipulation（2026-09-11）
  • 2609.05282 · Temporal Tactile Encoding and Compliance for Intent-Aware Robot-to-Human Bimanual Handover（2026-09-04）
  • 2609.05266 · TacPAC: Tactile Prediction and Real-Time Action Correction in World-Action Models for Contact-Rich Manipulation（2026-09-04）

可变形物体与材质状态数据

Section titled “可变形物体与材质状态数据”

7 篇。阅读主题文档。

  • 2609.17035 · SWIM: Vision-Language-Grounded Soft Whole-Body Interactive Manipulation（2026-09-15）
  • 2609.12677 · Size Doesn’t Matter: Material-State Reinforcement Learning for Excavator Transferable Soil Manipulation（2026-09-11）
  • 2609.12634 · Online Material Estimation for Conditioned Diffusion Policy in Shaping Deformable Linear Objects（2026-09-11）
  • 2609.12433 · FoldNet++: a Large-Scale Synthetic Dataset for Robotic T-Shirt Folding and Unfolding（2026-09-11）
  • 2609.10308 · Deformable Object Manipulation under Partial Observability via Real-Time Full-Shape Estimation（2026-09-09）
  • 2609.10243 · FolDeX: A Physical-World Benchmark for Long-Horizon Robotic Manipulation of Deformable Objects（2026-09-09）
  • 2608.28570 · ChainSplat: A Physics-Inspired Screw-Theoretic Model for Learning Deformable Linear Object Dynamics from Multi-View RGB Videos（2026-08-28）

灵巧手、抓取与双臂操作

Section titled “灵巧手、抓取与双臂操作”

12 篇。阅读主题文档。

  • 2609.28281 · BrickCraft-Duo: Efficient Dual-Arm Skill Learning and Refinement for Compositional Long-Horizon Assembly（2026-09-23）
  • 2609.24896 · Steerable and Reactive Grasping Through Modular Design with a Three-Point Interface（2026-09-21）
  • 2609.17172 · Fingers as Legs: Learning Self-Supported Locomotion and Manipulation with an Anthropomorphic Hand（2026-09-15）
  • 2609.16586 · ProxiDex: Learning Dynamics-Guided Proximity Policy for Dexterous Manipulation（2026-09-15）
  • 2609.16319 · ConGraspXL: Controllable Constraint-Conditioned Dexterous Grasping Motion Synthesis（2026-09-14）
  • 2609.14310 · VLBiMan++: Expanding the Generalization Boundary of Vision-Language Anchored One-Shot Bimanual Manipulation（2026-09-13）
  • 2609.13761 · Learning In-Hand Object Reaching to General 6D Poses（2026-09-12）
  • 2609.12498 · ArtManip: Category-Level Articulated In-Hand Manipulation（2026-09-11）
  • 2609.10137 · Assembling Two Parts in One Hand（2026-09-09）
  • 2609.05206 · Morphology and actuation as inductive biases in robotic hand manipulation（2026-09-04）
  • 2608.28578 · Aero Hand Open: A Simulation-Ready Tendon-Driven Hand for Dexterous Manipulation Learning（2026-08-28）
  • 2608.27221 · Tensegrity Continuum Robots Enable Task-Adaptive Morphologies for Cooperative Behaviors（2026-08-27）

人形/足式全身运动数据

Section titled “人形/足式全身运动数据”

21 篇。阅读主题文档。

  • 2609.28378 · ForgetMimic: Motion Unlearning for Reinforcement Learning Humanoid Control（2026-09-23）
  • 2609.26564 · Learning Air-Ground Motion Control with Temporal Mode Switching and Cross-Terrain Tracking（2026-09-22）
  • 2609.24840 · PredActor: Predictive Action Diffusion for Steerable Onboard Humanoid Control（2026-09-21）
  • 2609.22073 · Duty Factor Predicts Robust Constrained Quadrupedal Locomotion Across Gait Types（2026-09-18）
  • 2609.20566 · OmniMimic: Dynamics-completed Motion Augmentation for Multi-style Omnidirectional Quadruped Locomotion（2026-09-17）
  • 2609.20558 · Learning Slope-Adaptive Whole-Body Locomotion for Humanoid Robots in Roofing Construction（2026-09-17）
  • 2609.18930 · Learning Holistic Whole-Body Loco-Manipulation with a Bipedal Mobile Manipulator（2026-09-16）
  • 2609.18732 · PASSAGE: Scaling Scene-Aligned Motion Learning for Perceptive Humanoid Traversal in Cluttered Environments（2026-09-16）
  • 2609.16683 · Weave: Learning Whole-Body Dexterous Loco-Manipulation from Human-Object Interactions（2026-09-15）
  • 2609.16644 · WholeBodyWAM: Generalizing Pre-trained World-Action Priors to Humanoid Loco-Manipulation via WBC-Grounded Coordination（2026-09-15）
  • 2609.16405 · Collision-Aware Humanoid Whole-Body Control under Imperfect Tracking Targets（2026-09-14）
  • 2609.15770 · JEPLO: Joint-Embedding Predictive Learning for LiDAR-Based Legged Locomotion（2026-09-14）
  • 2609.15213 · X-WBC: A Cross-Embodiment Foundation Model for Humanoid Whole-Body Control（2026-09-14）
  • 2609.14647 · Skill Composition for Legged Robot Reinforcement Learning（2026-09-13）
  • 2609.14432 · EMoG: Emotion-Modulated Gait Generation for Expressive Humanoid Locomotion（2026-09-13）
  • 2609.12400 · Decentralized Evolution of Hexapod Gaits with Independent Leg Controllers（2026-09-11）
  • 2609.12347 · DWMP: Leveraging Dual World Models for Humanoid Obstacle Traversal（2026-09-11）
  • 2609.10286 · Learning Terrain-Adaptive Humanoid Locomotion on Granular Terrain（2026-09-09）
  • 2609.10283 · SwingBot: Learning Whole-Body Brachiation for Humanoid Robots（2026-09-09）
  • 2609.10273 · Frame-Coded Legged Locomotion over Noisy Terrain（2026-09-09）
  • 2609.03984 · MulDP: Multimodal Diffusion Policy for Autonomous Quadruped Parkour Navigation across Complex Terrains（2026-09-03）

导航、地图与可通行性数据

Section titled “导航、地图与可通行性数据”

10 篇。阅读主题文档。

  • 2609.21828 · Touvigation: Embodied Adaptive Object Acquisition for Blind and Low-Vision Users in Unfamiliar Indoor Environments（2026-09-18）
  • 2609.20604 · Semantic SLAM in Precision Agriculture using Bayesian Inference（2026-09-17）
  • 2609.18789 · AdaGeoVLN: Selective Geometry Across Representation Depth and Navigation Time for Vision-Language Navigation（2026-09-16）
  • 2609.17257 · Exploring 2D backbone effects for indoor semantic occupancy prediction（2026-09-15）
  • 2609.15195 · HarnessVLN: Unifying Training-Free Embodied Navigation through an Agent Harness（2026-09-14）
  • 2609.12422 · Hierarchical Belief Modeling for Zero-Shot Opponent Adaptation in Partially Observable Multi-Agent Navigation（2026-09-11）
  • 2609.06476 · One MLLM, One Call: Efficient Zero-Shot Vision-and-Language Navigation via Spatial-Aware Waypoints（2026-09-06）
  • 2609.06424 · OVMAN: A Task and Benchmark for Open-Vocabulary Motion-Aware Navigation（2026-09-06）
  • 2609.06251 · MobileVLA-R1 2.0: RL-Enhanced Reasoning for Mobile Robot Control（2026-09-05）
  • 2608.30935 · LightNav-0: Eliciting VLM Spatial Intelligence for Generalist Embodied Navigation（2026-08-31）

安全、鲁棒性、隐私与对抗评估

Section titled “安全、鲁棒性、隐私与对抗评估”

13 篇。阅读主题文档。

  • 2609.30258 · Temporal Gradient Inversion for Private Trajectory Reconstruction in Embodied Reinforcement Learning（2026-09-24）
  • 2609.22075 · LIMBO: Learning and Internalizing Model-Free Barrier Objectives for Agile and Safe Whole-Body Control（2026-09-18）
  • 2609.20822 · Coding Agents with an Obstacle-Aware Harness for Safe Robot Manipulation（2026-09-17）
  • 2609.17349 · RobResilience: Implementing and Evaluating a Resilience Framework for Cyber-Physical Embodied Systems（2026-09-15）
  • 2609.16724 · CorrRisk-WM: Corridor-Conditioned Risk World Modeling for Safety-Critical Trajectory Planning（2026-09-15）
  • 2609.15988 · ResSafe: Learning Safety Filtering with Residual Reinforcement Learning for Humanoids（2026-09-14）
  • 2609.15781 · When the World Lies: Backdoor Attacks on Latent World Models for Downstream Control（2026-09-14）
  • 2609.15113 · Legislating World-Model-Based Planning with Legal Reasoning（2026-09-14）
  • 2609.12371 · READ: Learning Risk-Informed Fields for End-to-End Autonomous Driving（2026-09-11）
  • 2609.10292 · Isotropic Embedding Perturbations for Robust Vision Language Encoders（2026-09-09）
  • 2609.06326 · Rethinking Safety for Generalist Robots（2026-09-06）
  • 2608.30428 · Lies We Can See: Joint Verbal and Non-Verbal Deception by VLM Agents in Embodied Social Interactions（2026-08-31）
  • 2608.28518 · When Robots Mishear Us: Mapping the Safety Risks of Voice-Controlled Embodied AI（2026-08-28）

基准、质检、成功判定与误差归因

Section titled “基准、质检、成功判定与误差归因”

23 篇。阅读主题文档。

  • 2609.28366 · AnchorReasoning: A Visual Grounding and Causal Reasoning Dataset in Long-Tail Autonomous Driving Scenarios（2026-09-23）
  • 2609.28236 · EmbodiedMemory-Bench: Benchmarking Embodied Memory for Long-Horizon Embodied Tasks（2026-09-23）
  • 2609.26567 · Beyond End-Task Success: How to Audit Visual Experience Retrieval in Robotics（2026-09-22）
  • 2609.24778 · H2RBench: A Real-to-Sim Benchmark for Evaluating Human-to-Robot Transfer（2026-09-21）
  • 2609.22055 · Benchmarking World Models for Continual Learning on Compositional Tasks（2026-09-18）
  • 2609.21909 · Beyond Kinematics: Benchmarking Simulation Fidelity for Muscle-Driven Imitation Learning（2026-09-18）
  • 2609.16745 · The Latent That Never Was: A Forensic Re-run of the CVAE Ablation in Action Chunking Transformer（2026-09-15）
  • 2609.16610 · EgoPathBench: Evaluating Zero-Shot Egocentric Waypoint Decision-Making in Vision-Language Models（2026-09-15）
  • 2609.16443 · The Neverwhere Visual Parkour Benchmark Suite（2026-09-14）
  • 2609.15940 · Beyond Single-Axis Testing: Paired Evaluation of Compound Robustness in Vision-Language-Action Policies（2026-09-14）
  • 2609.15726 · Bench2Dex: Benchmarking Visuo-Tactile Bimanual Dexterous Manipulation Across Dexterous Hands（2026-09-14）
  • 2609.15005 · IMPACT-VLA: Interaction-aware Multimodal Propagation Attribution via Counterfactual Trajectories for Vision-Language-Action Policies（2026-09-14）
  • 2609.14899 · What Makes a 3D Scene Editable? A Factorized Benchmark of Fidelity, Locality, Consistency, and Preservation（2026-09-14）
  • 2609.14833 · One Model, Two Physical Stories: Auditing Misalignment in Multi-Modal World Modeling（2026-09-13）
  • 2609.14473 · PuzzleMate: Benchmarking MLLMs for Egocentric Puzzle Assistance（2026-09-13）
  • 2609.13730 · JumpStart Your Policy Learning with Lessons from 160,000 Training Runs（2026-09-12）
  • 2609.13679 · How to Better Train VLAs: Lessons Learned From the REAL-I Challenge at ICRA 2026（2026-09-12）
  • 2609.13458 · STAGE: Diagnosing Semantic Transfer at Grounded Execution in Embodied Agents（2026-09-11）
  • 2609.13082 · Embodied-BenchForge: A Closed-Loop Agentic Workflow for Embodied Benchmark Construction（2026-09-11）
  • 2609.05324 · RoboSPA: Can VLA Models Go Beyond Simple Scenes and Short-Horizon Tasks?（2026-09-04）
  • 2609.02688 · From Proxy Learning to Driving Decisions: A Transfer-Based Framework for Evaluating Future-Aware Autonomous Driving Planners（2026-09-02）
  • 2608.30536 · Behavior-Skill: A Fine-Grained Benchmark for Evaluating Vision-Language-Action Policies in Long-Horizon Tasks（2026-08-31）
  • 2608.27345 · PAWBench: How Far Are We from Probabilistically Aligned World Modeling?（2026-08-27）

奖励、价值函数与强化学习反馈

Section titled “奖励、价值函数与强化学习反馈”

8 篇。阅读主题文档。

  • 2609.30023 · Res-HIL: Human-Guided Residual Reinforcement Learning for Sample-Efficient Dexterous Manipulation（2026-09-24）
  • 2609.22085 · SeeQ: Training Generalist Value Functions for Long-Horizon Robotic Manipulation（2026-09-18）
  • 2609.17115 · Intrinsic Robot Rewarding: Reusing VLA Representations for Autonomous Evaluation and Policy Improvement（2026-09-15）
  • 2609.15014 · Steering Generative Robot Policies with Lexicographic Preferences（2026-09-14）
  • 2609.14878 · Real-World Reinforcement Learning with MPC Scaffolding for Dexterous Manipulation（2026-09-14）
  • 2609.14261 · VGFM: Expressive Robot Policies via Dense Value Guidance in Flow Matching（2026-09-13）
  • 2609.12749 · SCQ: Stabilizing Conservative Q-Learning with Sigmoid-Bounded Entropy（2026-09-11）
  • 2608.30983 · Autonomously Acquiring Robot Manipulation Skills with Language-Driven Quality-Diversity（2026-08-31）

轨迹规划、控制与物理约束

Section titled “轨迹规划、控制与物理约束”

13 篇。阅读主题文档。

  • 2609.30213 · ReVAMP: Vector-Accelerated Motion Planning for Kinematically-Constrained Systems via Reparameterization（2026-09-24）
  • 2609.30140 · Contact as a Decision Variable: Capability-Tradeoff Contact Selection for Legged Loco-Manipulation（2026-09-24）
  • 2609.28377 · Amplify: A Lightweight Library for Reproducible Nonlinear Programming Problems in Robotics（2026-09-23）
  • 2609.28299 · Contact-Implicit Stein Projected ADMM for Discovery of Diverse Contact-Rich Manipulation Strategies（2026-09-23）
  • 2609.24841 · CAST: Collision-Aware Assembly with Construction Robots using Simultaneous Trajectory Estimation and Planning（2026-09-21）
  • 2609.22062 · Gripper-Aware Automatic Dense Packing of Irregular Objects（2026-09-18）
  • 2609.21803 · Contact-Rich Motion Planning via GPU-Parallel Mode Evaluation（2026-09-18）
  • 2609.18910 · CaSCo: Cascade-Aware Soft-Collision Motion Planning（2026-09-16）
  • 2609.15447 · Learning to Exploit Passive Dynamics for Energy-Efficient Target Hopping of a Spring-Legged Quadcopter（2026-09-14）
  • 2609.15399 · Dynamics-Informed Reinforcement Learning for Agile and Energy-Efficient Locomotion of a Monopedal Hopping Quadcopter（2026-09-14）
  • 2609.14868 · Primitive-Informed Sampling-Based MPC for Multi-Fingered Dexterous Manipulation（2026-09-14）
  • 2609.04070 · Continuous Actions from Discrete Minds: Latent-Aligned Planning for End-to-End Autonomous Driving（2026-09-03）
  • 2608.27186 · Task-space model-based control of pneumatic soft actuators（2026-08-27）

工业、农业、水下、医疗等特定域数据

Section titled “工业、农业、水下、医疗等特定域数据”

12 篇。阅读主题文档。

  • 2609.24761 · A Switched Adaptive Control Framework for Aerial Manipulators Under Dynamic Transitions（2026-09-21）
  • 2609.18881 · Body-Motion Control of a Simulated Aerial Swarm from a First-Person View（2026-09-16）
  • 2609.17021 · sensVLA: Spatially-Grounded Vision-Language-Action Model for Autonomous Wheel Loader（2026-09-15）
  • 2609.16880 · Artificial Intelligence-Enabled Space Robot Operations: Technologies, Challenges and Prospects（2026-09-15）
  • 2609.16696 · IL-ACT: Imitation Learning with Adaptive Cartesian Tracking Control for a 30-ton Excavator（2026-09-15）
  • 2609.16369 · Autonomous Droplet Navigation via Model-Based Reinforcement Learning（2026-09-14）
  • 2609.16186 · Occupancy Network-Guided Autonomous Robotic Partial Nephrectomy（2026-09-14）
  • 2609.15861 · DuctAM: A Duct-Assisted Quadrotor-Based Aerial Manipulator Enabling High-Force Push-and-Pull Interactions（2026-09-14）
  • 2609.14254 · Embracing Flow Unsteadiness: A High-Throughput Learning Platform Enables Vortex-Exploiting Bioinspired Propulsion（2026-09-13）
  • 2609.14198 · Novel Ex-vivo Calf Brain Model with Integrated Sub-Skull Force Sensors to Access Simulated Neurosurgical Procedures（2026-09-13）
  • 2609.12927 · Robust Underwater Grasping of Sloped Objects with a Waterproof Passive Adaptive Gripper（2026-09-11）
  • 2609.12883 · From Transportation to Manipulation: Enabling Grasping in Magnetic Robotics（2026-09-11）

覆盖审计

Section titled “覆盖审计”

下图按主主题统计每页所关联的论文数量。主题大小代表文献记录量，不代表研究质量或数据集价值；少量高质量、可复现的真实数据仍可能比大量不完整数据更有用。

26 个独立知识点主题的论文数量分布及 36 篇范围外记录

暂未纳入的论文

Section titled “暂未纳入的论文”

以下 36 篇被关键词检索命中，但原始摘要未直接支撑本站的数据采集、处理、标注、评估或数据集使用主题。此处仅说明文档范围，不评价论文质量；边界案例可在获得更直接的数据证据后重新归类。

  • 2609.28416 · Agent-Editing World Model: Rethinking World Modeling for LLM Agents（2026-09-23）
  • 2609.28335 · An Open Pipeline and Dashboard for Systemic-Risk Evidence under the EU AI Act’s Code of Practice（2026-09-23）
  • 2609.28182 · Finite-Sample Probabilistic Safety Certification for AI-Based Grid-Edge Coordination（2026-09-23）
  • 2609.26761 · A2M: Trace-Optimized Agent Hijacking in the MCP Ecosystem（2026-09-22）
  • 2609.24801 · Decoding Guardrails: XAI-Guided Perturbation Analysis of Prompt Injection Detection（2026-09-21）
  • 2609.24742 · LLM-based Conversational AI Knowledge Assistant for MyBuddy Humanoid Robot（2026-09-21）
  • 2609.21906 · Intervention Granularity Matters: Coherent Treatment Bundles in Counterfactual Simulation with Clinical World Models（2026-09-18）
  • 2609.17325 · Intrinsic Motivation in Reinforcement Learning: A Research Agenda for Adaptive Self-Organisation（2026-09-15）
  • 2609.17221 · Grounding SWE-Agent Decisions in Architecture-0 Design: Navigating Unknown Unknowns through Physical Mapping（2026-09-15）
  • 2609.16778 · Unifying Semantic Priors and High-Frequency Traces: Enhancing V-JEPA with Mixture-of-Experts for Robust Synthetic Image Forensics（2026-09-15）
  • 2609.16679 · AI for Games in the Foundation Model Era（2026-09-15）
  • 2609.16436 · Interpreting and Steering LLM Agents for Social Simulations（2026-09-14）
  • 2609.16409 · Reasoning with Image Generation（2026-09-14）
  • 2609.15975 · Disentangling Representation Evolution in Transformers through Directional Decomposition（2026-09-14）
  • 2609.16098 · Universal Defenses for Tool-Integrated LLM Agents Against Adversarial Attacks（2026-09-14）
  • 2609.15277 · Artificial entrepreneurial cognition: Locating and causally steering an opportunity recognition dial inside large language models (LLMs)（2026-09-14）
  • 2609.15234 · CWM: Controllable White-Box Meta-Prompting for Adaptive Retrieval-Augmented Generation and Reasoning Ability（2026-09-14）
  • 2609.14709 · An immune world model for multiscale forecasting and therapeutic hypothesis generation（2026-09-13）
  • 2609.14437 · Lightweight Generalized DeepFake Face Detection with WAVIE: Wavelet Augmented Vision Intermediate Embeddings（2026-09-13）
  • 2609.16062 · Digital Persuasion: Understanding the Impact of Online Influencers on Public Opinion（2026-09-13）
  • 2609.13927 · Exploring napping paradigm for Recurrent Spiking Neural Networks（2026-09-12）
  • 2609.13771 · Homeostatic Continual Learning（2026-09-12）
  • 2609.13499 · Canaries in the Bank: Auditing User-Level Privacy in Private Evolution（2026-09-11）
  • 2609.13352 · Real-time Learning and Evolution in Robotic Art Installations（2026-09-11）
  • 2609.12668 · Beyond Ambiguous Visual Cues: Studying Physiological Disruptions and Cross-Modal Inconsistencies in Deepfake Videos（2026-09-11）
  • 2609.12606 · Beyond Generation and Accuracy: Diagnosing and Enhancing Visual Chain-of-Thought for Geometry Problem Solving（2026-09-11）
  • 2609.10372 · PACE: Perceived-Latency-Aware Cascading Service Routing and Filler Control for QoE-Efficient Retrieval-Augmented Dialogue Serving（2026-09-09）
  • 2609.06530 · Selective Knowledge Control for Continual GUI Agent Learning over Application Streams（2026-09-06）
  • 2609.05404 · Diffusion TV: Experiencing Diffusion Models through Tangible, Embodied Interaction（2026-09-04）
  • 2609.05227 · CABAL: Multi-Agent Simulacra for Tracing the Effects of Collusive Bidding in Peer Review（2026-09-04）
  • 2609.02885 · Discriminative World Models for Web Agents（2026-09-02）
  • 2609.02797 · Dutch Books for Language Models（2026-09-02）
  • 2609.02640 · From Detection to Localization: A Unified Forensics Framework for Fully Synthetic and Tampered Images（2026-09-02）
  • 2608.30656 · APT: Anchor-aligned Perturbations for Tamper Localization in Fully Regenerated Images（2026-08-31）
  • 2608.28541 · An Enclosed Mode Is a Gauge Choice: Topology Relative to Reach in Certified Code World Models（2026-08-28）
  • 2608.28302 · FUSED: Forensic-Semantic Mixture-of-Experts for AI Inpainting Detection and Localization（2026-08-28）

来源与更新口径

Section titled “来源与更新口径”
  • 原始记录：spiders/arxiv/papers/2026-*.json，以单篇 published 和 arXiv ID 为准；latest.json 是滚动窗口，不作为全月来源。
  • 主题判断：按具身智能数据用途归类。纯通用 Web/GUI/LLM 代理、图像取证、非机器人临床/金融/电网研究等不因关键词命中就进入正文。
  • 复核方法：可从本页每条 arXiv 链接核对标题、摘要、版本和论文正文。数字结论只应在对应论文的实验设置下解释。