Embodied AI Radar · Weekly 2026-W42 · 具身智能雷达周报
2026-10-12 → 2026-10-18 · 382 new arXiv papers · 20 with code
本周共收录 382 篇 arXiv 新论文，其中 20 篇附代码。
This is the first digest: the radar launched on 2026-10-19 with a backfill of this week's papers. CoRL 2026 acceptances dominate the venue list below.
这是第一期周报：雷达于 2026-10-19 上线，并回填了本周论文。下方会议接收列表以 CoRL 2026 为主。
Editor's picks · 本周精选
- The first agent picks from this week's papers were added on launch day and appear in next week's digest and on the home page. 本周论文中的首批精选于上线当天加入，见下周周报与网站首页。
Accepted at venues · 已被会议/期刊接收
From the authors' arXiv comments. 根据作者在 arXiv 上的备注。
- CoRL 2026 · PathLift: Depth-Conditioned Metric Lifting for Visual-Navigation Priors
- CoRL 2026 · FlowVLA: Enabling World Dynamics Understanding through Pluggable 3D Flow Learning
- CoRL 2026 · PACE: Temporal-Role Tactile Conditioning for Chunk-Based Insertion Policies · project
- CoRL 2026 · Constraint-Decomposed Impedance Control from Generative Policy for Articulated Object Manipulation · project
- CoRL 2026 · SHOES: Sim-to-real Hardware-adaptation via On-policy Expansion of Skills for Legged Robots
- CoRL 2026 · DeTac: Dual-branch Decoupling of Texture and Force Signals in Vision-Based Tactile Sensors
- CoRL 2026 · Phase-Conditioned Affordance Fields for SE(3)-Equivariant Visuomotor Policies · project
- CoRL 2026 · Training Generalist Navigation Policies via Action Self-Distillation
- CoRL 2026 · HelpMeOut: A Neuropsychology-Grounded Benchmark for Error-Aware Monitoring of Household Task Execution · project
- CoRL 2026 · Where Small Mistakes Matter: Quantifying State Criticality for Imitation Learning · project
- CoRL 2026 · Beyond the End-Effector: Learning Whole-Arm Contact Manipulation from Human Demonstrations · project
- CoRL 2026 · GLST: Global-Local Spatio-Temporal Memory Decomposition Framework for Online Chained-Goal Navigation
- CoRL 2026 · Walk Like This: Stylized Locomotion for Physics-Based Characters through Example-Driven Diffusion
- CoRL 2026 · CoSMO: Conformal Semantic Mapping for Open-World Safe Robot Navigation · project
- CoRL 2026 · Active Generative Simulation: Meta-Learning What to Simulate for Sim-and-Real Policy Co-Training · project
- CoRL 2026 · Graph-Reweighted Behavior Cloning with Human Interventions for Precise Robotic Manipulation · project
- CoRL 2026 · Hierarchical Residual Policy Learning for Real-World Mobile Manipulation with Sparse Human Guidance · project
- CoRL 2026 · SafeDriveVLA: Navigation-Conditioned World Model Dreaming for Conflict-Aware End-to-End Autonomous Driving · project
- CoRL 2026 · An Agentic Framework for Proactive Robots with Procedural Assistance Memory · project
- CoRL 2026 · Asking Smarter Questions: Targeted Active Preference Learning for Underspecified Rewards
- IROS 2026 · All-St3R SLAM: Real-time RGB Dynamic SLAM with Pixel-wise Motion Masking · project
- IROS 2026 · MSC: Multi-Stage Caster-aware Control Framework with Uncertainty-Aware NMPC and RL-based Execution Layer
- IROS 2026 · Improving Policy Learning from Delayed Interventions
- IROS 2026 · Collective Learning for Unified Robotic Manipulation via Predictive Dynamics Adaptation
- IROS 2026 · MR-MPCC: Model Predictive Contouring Control for Robotic Manipulator Tracking under Moving Reference Frames
- IROS 2026 · SPEAR: Selective Preference Elicitation with Adaptive Rewards for VLM-Guided Robotic Manipulation · project
- NeurIPS 2026 · Efficient Test-Time Adaptation For Robot Policies
- NeurIPS 2026 · LIBERO-PeRM: Benchmarking Personalized Robotic Manipulation · project
- RSS 2026 · Perception Beyond the Visible: Incompleteness-Aware Scene Understanding for Robot Autonomy · project
- WACV 2027 · FastVLA: Task-Conditioned and Temporally Adaptive Visual Token Compression for Efficient Vision-Language-Action Models · code
By topic · 分主题统计
- VLA & Robot Foundation Models · VLA 与机器人基础模型: 51
- Manipulation & Dexterous Hands · 操作与灵巧手: 77
- Tactile & Visuotactile Sensing · 触觉与视触觉感知: 16
- Locomotion & Humanoids · 运动控制与人形机器人: 29
- World Models · 世界模型: 34
- Navigation & Embodied Agents · 导航与具身智能体: 46
- Perception, SLAM & 3D Mapping · 感知、SLAM 与三维建图: 18
- Planning, Control & Safety · 规划、控制与安全: 23
- Sim-to-Real & Simulators · 仿真与 Sim-to-Real: 14
- Datasets & Benchmarks · 数据集与基准: 26
- Hardware & Open-Source Robots · 硬件与开源机器人: 1
New papers with code · 本周附代码的新论文
VLA & Robot Foundation Models · VLA 与机器人基础模型
- REKI: Registering Knowledge in VLAs · code
- FastVLA: Task-Conditioned and Temporally Adaptive Visual Token Compression for Efficient Vision-Language-Action Models · code
- LTM: Latent Transition Based In-Context Memory for Vision-Language-Action Models · code
- VLA-$Δ$: Learning Physical State Transition for Vision-Language-Action Models · code
- Cracking Mobile Manipulation: From Failure Diagnosis to Agentic Robot System · code
- Embodied Skill Bench: Benchmarking How Well Agents with Embodied Skills Perform on Physical Tasks · code
Manipulation & Dexterous Hands · 操作与灵巧手
- Cracking Mobile Manipulation: From Failure Diagnosis to Agentic Robot System · code
- Can a Frontier VLM Serve as a Robot Policy for Memory-Dependent Manipulation? · code · project
Navigation & Embodied Agents · 导航与具身智能体
Planning, Control & Safety · 规划、控制与安全
Datasets & Benchmarks · 数据集与基准
- Embodied Skill Bench: Benchmarking How Well Agents with Embodied Skills Perform on Physical Tasks · code
---
Auto-tagged by keyword rules; picks are reviewed by the maintainer agent. Website · GitHub
主题为关键词规则自动标注；精选由维护智能体审阅。