机器人论文速览

Research domain

具身大模型

177 papers · latest 2026-07-17

具身大模型

Jul 17, 2026Rec 3/5Score 82

[AC-VLA] AC-VLA: Robust Out-of-Distribution Action Execution via Compositional Learning

[AC-VLA] AC-VLA：通过组合学习实现鲁棒分布外动作执行

VLA模型用于机器人动作执行，直接涉及具身操作

paper

具身大模型

Jul 17, 2026Rec 3/5Score 75

Vision-Language-Motion Maps: An Open-Vocabulary, Uncertainty-Aware, Queryable Motion Attribute for 3D Scene Maps

视觉-语言-运动地图: 开放词汇可查询的3D场景运动属性

VLM与运动规划结合，支持开放词汇查询3D场景

paper

具身大模型

Jul 17, 2026Rec 3/5Score 73

[JoyNexus] JoyNexus: Service-Oriented Multi-Tenant Post-Training for VLA Models

JoyNexus: 服务导向的多租户VLA模型后训练

VLA模型服务机器人后训练，提升多任务能力

paper

具身大模型

Jul 10, 2026Rec 3/5Score 75

[en] Communication-Efficient Digital-Twin Coordination for Heterogeneous LLM Embodied Agents over Computing Power Networks

算力网络上异构LLM具身智能体的通信高效数字孪生协调

LLM具身智能体的多智能体协调方法

paper

Research domain

具身操作

438 papers · latest 2026-07-17

具身操作

Jul 17, 2026Rec 3/5Score 85

[Humanoid] Handroid: Bridging Dexterous Hand and Humanoid

[Humanoid] Handroid：连接灵巧手与人形机器人

灵巧手与人形结合，涉及具身操作技能

paper

具身操作

Jul 17, 2026Rec 3/5Score 78

Dynamics-Aware Meta-Imitation for Generalization to Unseen Robotic Manipulation

动态感知元模仿学习用于泛化到未知机器人操作

元模仿学习方法提升操作泛化能力，含Sim-to-Real

paper

具身操作

Jul 17, 2026Rec 3/5Score 70

Data and Learning Where it Matters for Contact-Rich Manipulation

接触丰富操作的数据与学习关键问题研究

研究接触丰富操作的数据效率和Sim-to-Real问题

paper

具身操作

Jul 10, 2026Rec 3/5Score 82

[en] Tactile and Vision Conditioned Contact-Centric Control for Whole-Arm Manipulation

基于触觉和视觉的接触中心控制实现整臂操作

触视觉融合控制整臂操作，具身操作核心

paper

Research domain

具身导航

197 papers · latest 2026-07-17

具身导航

Jul 17, 2026Rec 3/5Score 88

[RAVEN] RAVEN: Reinforcement-Adaptive Visibility-Graph Planning for Robust Humanoid Navigation with Collision-Free MPC

[RAVEN] RAVEN：基于强化自适应可见性图规划的无碰撞人形机器人鲁棒导航

人形机器人导航与MPC控制，直接涉及具身导航

paper

具身导航

Jul 17, 2026Rec 3/5Score 78

[SkillNav] SkillNav: Score-Level Skill Intervention for Zero-Shot Object Goal Navigation

SkillNav：零样本目标导航的分数级技能干预

零样本导航技术结合技能干预，适用于服务机器人

paper

具身导航

Jul 17, 2026Rec 3/5Score 72

Embodied Active Learning under Limited Annotation and Navigation Budget for Object Detection

有限标注与导航预算下的具身主动学习用于目标检测

具身主动学习用于导航中的目标检测，契合具身导航

paper

具身导航

Jul 17, 2026Rec 3/5Score 72

Difference-Based Relational Learning for Zero-Shot Object-Goal Visual Navigation With Direct Sim-to-Real Transfer

基于差异的关系学习用于零样本目标视觉导航直传Sim-to-Real

零样本视觉导航方法，支持Sim-to-Real直接迁移

paper

Research domain

强化学习

153 papers · latest 2026-07-17

强化学习

Jul 17, 2026Rec 3/5Score 65

When Does Muon Help Agentic Reinforcement Learning?

Muon何时助力代理强化学习

代理强化学习与具身智能相关

paper

强化学习

Jul 10, 2026Rec 3/5Score 65

[PAC-ACT] PAC-ACT: Post-training Actor-Critic for Action Chunking Transformers

PAC-ACT：动作分块Transformer的后训练Actor-Critic方法

强化学习动作控制研究，应用于机器人操作

paper

强化学习

Jul 10, 2026Rec 3/5Score 65

Multimodal Reward Hacking in Reinforcement Learning

强化学习中的多模态奖励黑客

多模态奖励设计对机器人RL学习有重要意义

paper

强化学习

Jul 10, 2026Rec 3/5Score 60

[CORAL-AUV] CORAL-AUV: CFD Oriented Reinforcement Learning for Autonomous Underwater Vehicles

CORAL-AUV：面向自主水下航行器的CFD导向强化学习

水下航行器是具身智能的一种应用形态

paper

Research domain

仿真平台

23 papers · latest 2026-07-10

仿真平台

Jul 10, 2026Rec 3/5Score 60

[DemoBridge] DemoBridge: A Simulation-in-the-Loop Toolkit for Single-View Human Demonstration Retargeting

DemoBridge：单视角人类演示重定向的仿真工具包

人机演示迁移仿真平台，支撑具身操作

paper

仿真平台

Jul 2, 2026Rec 3/5Score 88

[Embodied.cpp] Embodied.cpp: A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots

Embodied.cpp：异构机器人具身AI模型可移植推理运行时

具身AI模型推理运行时，支持异构机器人部署

paper

仿真平台

Jul 2, 2026Rec 3/5Score 75

Controllable Sim Agents with Behavior Latents

基于行为潜变量的可控模拟智能体

仿真智能体与具身AI行为控制高度相关

paper

仿真平台

Jun 26, 2026Rec 3/5Score 82

[en] Building a Scalable, Reproducible, Evaluatable, and Closed-Loop Simulation Environment Foundation for Embodied Intelligence Cloud-Native Simulation Infrastructure for Embodied Intelligence Training, Evaluation, and Data Collection

具身智能云原生仿真基础设施

构建具身智能仿真平台

paper

Research domain

环境感知

202 papers · latest 2026-07-17

环境感知

Jul 17, 2026Rec 3/5Score 80

[Hyper-Dexterous] Towards Artificial Nerves: Biomimetic Optical-Fiber Tactile Sensing for Robots

迈向人工神经：用于机器人的仿生光纤触觉传感

机器人仿生触觉传感直接相关

paper

环境感知

Jul 17, 2026Rec 3/5Score 70

A Model-Based Decoupling Strategy for Proprioception and Contact Sensing in an Architected Soft Manipulator

架构化软体机械臂的本体感与接触感理解耦策略

软体机械臂感知与操控相关

paper

环境感知

Jul 17, 2026Rec 3/5Score 68

[VTLoc] VTLoc: Learning-based Tactile Contact Localization in Visual Point Clouds

VTLoc: 基于学习的视觉点云触觉接触定位

触觉感知与点云结合用于接触定位的灵巧操作

paper

环境感知

Jul 17, 2026Rec 3/5Score 55

An Exam for Active Observers

主动观察者考试

主动观察者评估涉及具身感知

paper

Research domain

任务规划

89 papers · latest 2026-07-17

任务规划

Jul 17, 2026Rec 3/5Score 65

A Generative Partially Specified Finite State Machine Approach to Complex Behaviour Planning

生成式部分指定有限状态机方法用于复杂行为规划

有限状态机结合生成模型进行任务规划，具身应用潜力

paper

任务规划

Jul 10, 2026Rec 3/5Score 70

[en] GenVid2Robot: From Video Generation to Robot Manipulation via Rigid-Geometric Consistency

GenVid2Robot：通过刚性几何一致性从视频生成到机器人操作

视频生成引导机器人操作的任务规划

paper

任务规划

Jul 10, 2026Rec 3/5Score 65

Task-Adaptive Design of Modular Aerial Manipulators Under Airflow Exposure Constraints

气流暴露约束下模块化空中机械手的任务自适应设计

空中机械手的任务与运动规划

paper

任务规划

Jun 18, 2026Rec 3/5Score 75

Hierarchical Control in Multi-Agent Games: LLM-based Planning and RL Execution

多智能体博弈分层控制：基于LLM的规划和RL执行

多智能体任务规划与LLM应用

paper

Research domain

具身交互

136 papers · latest 2026-07-17

具身交互

Jul 17, 2026Rec 3/5Score 82

[Let the Body Follow] Let the Body Follow: Coupled Egocentric Control for Whole-Body Robot Teleoperation

Let the Body Follow：全身机器人遥操作的耦合自我中心控制

全身遥控制结合自我中心控制，核心具身交互研究

paper

具身交互

Jul 17, 2026Rec 3/5Score 72

[Humanoid] PACE: Persona Adaptation through Conversational Elicitation in Human-Robot Interaction

PACE：人机交互中的对话启发式人格适配

人形机器人对话交互人格适配，具身大模型应用

paper

具身交互

Jul 10, 2026Rec 3/5Score 78

[en] Effects of Robotic Touch on Older Users During Walking Guidance by a Humanoid Robot

人形机器人行走引导中机械触摸对老年用户的影响

人形机器人触觉交互与行走引导直接相关

paper

具身交互

Jul 10, 2026Rec 3/5Score 75

[en] Validating Virtual Reality for Studying Multimodal Human-Robot Interaction in Socially Aware Robot Navigation

验证虚拟现实用于社会感知机器人导航的多模态人机交互研究

VR研究社会导航中的人机交互，高度相关

paper

Research domain

运动规划

284 papers · latest 2026-07-17

运动规划

Jul 17, 2026Rec 3/5Score 68

A Task-Space Receding Horizon Controller for Fast Collision Avoidance

用于快速避障的任务空间后退地平线控制器

任务空间MPC控制器用于避障，具身运动规划核心

paper

运动规划

Jul 10, 2026Rec 3/5Score 68

[en] Shortcut Trajectory Planning for Efficient Offline Reinforcement Learning

高效离线强化学习的捷径轨迹规划

离线强化学习轨迹规划，服务具身智能决策

paper

运动规划

Jul 2, 2026Rec 3/5Score 85

[The Moving Eye] The Moving Eye: Enhancing VLA Spatial Generalization via Hybrid Dynamic Data Collection

[The Moving Eye] Moving Eye：通过混合动态数据采集增强VLA空间泛化

VLA空间泛化能力提升，具身操作核心研究方向

paper

运动规划

Jul 2, 2026Rec 3/5Score 82

[VLA-Corrector] VLA-Corrector: Lightweight Detect-and-Correct Inference for Adaptive Action Horizon

[VLA-Corrector] VLA-Corrector：自适应动作视野轻量级检测修正

VLA模型推理修正技术，直接服务具身动作规划

paper

Research domain

机器人安全

247 papers · latest 2026-07-10

机器人安全

Jul 10, 2026Rec 3/5Score 60

[en] Empirical Pedestrian Safety Assessment in a Mobile Robot Using a Predictive Social Force Model

基于预测社会力模型的移动机器人行人安全实证评估

移动机器人安全导航与行人交互

paper

机器人安全

Jul 2, 2026Rec 3/5Score 60

Hardware-Enforced Semantic Coordination for Safety-Critical Real-Time Autonomous Systems

安全关键实时自主系统的硬件强制语义协调

自动驾驶安全系统直接具身相关

paper

机器人安全

Jun 26, 2026Rec 3/5Score 72

[en] PPO-EAL: Exact Augmented Lagrangian Proximal Policy Optimization for Safe Robotic Control

PPO-EAL：精确增广拉格朗日近端策略优化安全机器人控制

PPO强化学习安全控制方法直接用于机器人

paper

机器人安全

Jun 12, 2026Rec 3/5Score 78

Provably Safe, Yet Scalable Reinforcement Learning

可证明安全且可扩展的强化学习

在保证安全性的同时实现强化学习的可扩展性证明

paper

Research domain

群体智能

53 papers · latest 2026-07-17

群体智能

Jul 17, 2026Rec 3/5Score 60

[en] Implicit Virtual Leader: Decentralized Vision-Only Relative Pose Estimation for Multi-Robot Formations

隐式虚拟领导者：多机器人编队的去中心化纯视觉相对位姿估计

多机器人编队感知相关

paper

群体智能

Jul 10, 2026Rec 3/5Score 70

[CoDiMAD] CoDiMAD: Diffusion-Based Privileged Distillation for Communication-Free Multi-Robot Coordination

CoDiMAD：无通信多机器人协调的扩散式特权蒸馏

多机器人协调是群体智能核心问题

paper

群体智能

Jul 2, 2026Rec 3/5Score 55

[QFedAgent] QFedAgent: Quantum-Enhanced Personalized Federated Learning for Multi-Agent Activity Recognition

[QFedAgent] 量子增强个性化联邦学习多智能体活动识别

多智能体协同学习，群体智能与具身相关

paper

群体智能

Jun 26, 2026Rec 3/5Score 88

[en] When Multi-Robot Systems Meet Agentic AI:Towards Embodied Collective Intelligence

多机器人系统遇上智能体AI：迈向具身集体智能

多机器人具身智能研究直接符合具身智能领域

paper

Research domain

数据集评测

30 papers · latest 2026-07-17

数据集评测

Jul 17, 2026Rec 3/5Score 82

[Hyper-Dexterous] IMBench: A Benchmark for Intuitive Robotic Manipulation

IMBench: 直观机器人操作的基准测试

机器人操作基准测试，评估灵巧手和操作能力

paper

数据集评测

Jul 10, 2026Rec 3/5Score 85

[en] TactiDex: A Real-World Tactile-Guided Benchmark for Human-Like Dexterous Manipulation

TactiDex：面向类人灵巧操作的真实世界触觉引导基准

灵巧手操作基准数据集，触觉感知引导具身操作

paper

数据集评测

Jun 18, 2026Rec 3/5Score 55

[SARLO-80] SARLO-80: Worldwide Slant SAR Language Optic Dataset 80cm

[SARLO-80] SARLO-80: 全球倾斜SAR语言光学数据集80cm

SAR数据集可服务于具身感知

paper

数据集评测

Jun 12, 2026Rec 3/5Score 65

[Kine2Go] Kine2Go: Kinematic dataset for the Unitree Go2 robot with diverse gaits and motions

[Kine2Go] Kine2Go：Unitree Go2机器人多步态运动学数据集

发布Unitree Go2机器人的多样化步态运动学数据集

paper

Research domain

世界模型

22 papers · latest 2026-07-17

世界模型

Jul 17, 2026Rec 3/5Score 78

[Orbis 2] Orbis 2: A Hierarchical World Model for Driving

[Orbis 2] Orbis 2：一种用于驾驶的分层世界模型

世界模型用于自动驾驶场景，与具身导航相关

paper

世界模型

Jul 10, 2026Rec 3/5Score 82

[en] Causally Debiased Latent Action Model for Embodied Action Conditioned World Models

面向具身动作条件世界模型的因果无偏潜动作模型

具身智能世界模型构建，含因果推理与动作建模

paper

世界模型

Jul 2, 2026Rec 3/5Score 88

[PhysMani] PhysMani: Physics-principled 3D World Model for Dynamic Object Manipulation

PhysMani：面向动态物体操作的物理原则3D世界模型

物理启发的3D世界模型用于动态物体操作

paper

世界模型

Jul 2, 2026Rec 3/5Score 83

[ACID] ACID: Action Consistency via Inverse Dynamics for Planning with World Models

ACID：通过逆动力学实现动作一致性用于世界模型规划

逆动力学驱动的世界模型规划方法

paper

Research domain

Sim-to-Real

10 papers · latest 2026-07-02

Sim-to-Real

Jul 2, 2026Rec 3/5Score 78

[Learning] Actuator Reality Shaping for Zero-Shot Sim-to-Real Robot Learning

面向零样本Sim-to-Real机器人学习的执行器现实塑形

Sim-to-Real核心问题直接具身

paper

Sim-to-Real

Jun 18, 2026Rec 3/5Score 70

[en] Efficiently Linking Real Scenes with Synthetic Data Generation for AI-based Cognitive Robotics and Computer Vision Applications

高效关联真实场景与合成数据生成用于AI认知机器人和计算机视觉

合成数据生成支撑Sim-to-Real研究

paper

Sim-to-Real

Apr 10, 2026Rec 3/5Score 56

Sim-to-Real Transfer for Muscle-Actuated Robots via Generalized Actuator Networks

基于广义执行器网络的肌肉驱动机器人Sim-to-Real迁移

提出肌肉驱动机器人的广义执行器网络Sim-to-Real迁移

paper

Sim-to-Real

Nov 6, 2025Rec 3/5Score 56

Can Context Bridge the Reality Gap? Sim-to-Real Transfer of Context-Aware Policies

上下文能否弥合现实差距？上下文感知策略的Sim-to-Real迁移

研究上下文感知策略从模拟环境到真实世界的迁移效果

paper

Research domain

基础理论

7 papers · latest 2026-07-10

基础理论

Jul 10, 2026Rec 3/5Score 52

Robustifying Vision-Language Models via Test-Time Prompt Adaptation

通过测试时提示适应增强视觉语言模型

VLM鲁棒性提升技术，可支撑具身智能感知

paper

基础理论

Jun 12, 2026Rec 3/5Score 50

[HPSv3++] HPSv3++: Scaling Reward Models Across the Full Spectrum of Diffusion Model Capabilities

HPSv3++: 跨扩散模型能力全谱扩展奖励模型

扩展奖励模型覆盖扩散模型的全部能力范围

paper

基础理论

Dec 4, 2025Rec 3/5Score 54

When Robots Should Say "I Don't Know": Benchmarking Abstention in Embodied Question Answering

机器人何时应该说“我不知道”：具身问答中的放弃回答基准测试

提出具身问答中机器人放弃回答的评测基准

paper

基础理论

Oct 20, 2025Rec 3/5Score 54

[en] RoboChallenge: Large-scale Real-robot Evaluation of Embodied Policies

RoboChallenge：大规模真实机器人具身策略评估

提出大规模真实机器人具身策略评估方案

paper

tabris.cloud

机器人论文速览

具身大模型

[AC-VLA] AC-VLA: Robust Out-of-Distribution Action Execution via Compositional Learning

Vision-Language-Motion Maps: An Open-Vocabulary, Uncertainty-Aware, Queryable Motion Attribute for 3D Scene Maps

[JoyNexus] JoyNexus: Service-Oriented Multi-Tenant Post-Training for VLA Models

[en] Communication-Efficient Digital-Twin Coordination for Heterogeneous LLM Embodied Agents over Computing Power Networks

具身操作

[Humanoid] Handroid: Bridging Dexterous Hand and Humanoid

Dynamics-Aware Meta-Imitation for Generalization to Unseen Robotic Manipulation

Data and Learning Where it Matters for Contact-Rich Manipulation

[en] Tactile and Vision Conditioned Contact-Centric Control for Whole-Arm Manipulation

具身导航

[RAVEN] RAVEN: Reinforcement-Adaptive Visibility-Graph Planning for Robust Humanoid Navigation with Collision-Free MPC

[SkillNav] SkillNav: Score-Level Skill Intervention for Zero-Shot Object Goal Navigation

Embodied Active Learning under Limited Annotation and Navigation Budget for Object Detection

Difference-Based Relational Learning for Zero-Shot Object-Goal Visual Navigation With Direct Sim-to-Real Transfer

强化学习

When Does Muon Help Agentic Reinforcement Learning?

[PAC-ACT] PAC-ACT: Post-training Actor-Critic for Action Chunking Transformers

Multimodal Reward Hacking in Reinforcement Learning

[CORAL-AUV] CORAL-AUV: CFD Oriented Reinforcement Learning for Autonomous Underwater Vehicles

仿真平台

[DemoBridge] DemoBridge: A Simulation-in-the-Loop Toolkit for Single-View Human Demonstration Retargeting

[Embodied.cpp] Embodied.cpp: A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots

Controllable Sim Agents with Behavior Latents

[en] Building a Scalable, Reproducible, Evaluatable, and Closed-Loop Simulation Environment Foundation for Embodied Intelligence Cloud-Native Simulation Infrastructure for Embodied Intelligence Training, Evaluation, and Data Collection

环境感知

[Hyper-Dexterous] Towards Artificial Nerves: Biomimetic Optical-Fiber Tactile Sensing for Robots

A Model-Based Decoupling Strategy for Proprioception and Contact Sensing in an Architected Soft Manipulator

[VTLoc] VTLoc: Learning-based Tactile Contact Localization in Visual Point Clouds

An Exam for Active Observers

任务规划

A Generative Partially Specified Finite State Machine Approach to Complex Behaviour Planning

[en] GenVid2Robot: From Video Generation to Robot Manipulation via Rigid-Geometric Consistency

Task-Adaptive Design of Modular Aerial Manipulators Under Airflow Exposure Constraints

Hierarchical Control in Multi-Agent Games: LLM-based Planning and RL Execution

具身交互

[Let the Body Follow] Let the Body Follow: Coupled Egocentric Control for Whole-Body Robot Teleoperation

[Humanoid] PACE: Persona Adaptation through Conversational Elicitation in Human-Robot Interaction

[en] Effects of Robotic Touch on Older Users During Walking Guidance by a Humanoid Robot

[en] Validating Virtual Reality for Studying Multimodal Human-Robot Interaction in Socially Aware Robot Navigation

运动规划

A Task-Space Receding Horizon Controller for Fast Collision Avoidance

[en] Shortcut Trajectory Planning for Efficient Offline Reinforcement Learning

[The Moving Eye] The Moving Eye: Enhancing VLA Spatial Generalization via Hybrid Dynamic Data Collection

[VLA-Corrector] VLA-Corrector: Lightweight Detect-and-Correct Inference for Adaptive Action Horizon

机器人安全

[en] Empirical Pedestrian Safety Assessment in a Mobile Robot Using a Predictive Social Force Model

Hardware-Enforced Semantic Coordination for Safety-Critical Real-Time Autonomous Systems

[en] PPO-EAL: Exact Augmented Lagrangian Proximal Policy Optimization for Safe Robotic Control

Provably Safe, Yet Scalable Reinforcement Learning

群体智能

[en] Implicit Virtual Leader: Decentralized Vision-Only Relative Pose Estimation for Multi-Robot Formations

[CoDiMAD] CoDiMAD: Diffusion-Based Privileged Distillation for Communication-Free Multi-Robot Coordination

[QFedAgent] QFedAgent: Quantum-Enhanced Personalized Federated Learning for Multi-Agent Activity Recognition

[en] When Multi-Robot Systems Meet Agentic AI:Towards Embodied Collective Intelligence

数据集评测

[Hyper-Dexterous] IMBench: A Benchmark for Intuitive Robotic Manipulation

[en] TactiDex: A Real-World Tactile-Guided Benchmark for Human-Like Dexterous Manipulation

[SARLO-80] SARLO-80: Worldwide Slant SAR Language Optic Dataset 80cm

[Kine2Go] Kine2Go: Kinematic dataset for the Unitree Go2 robot with diverse gaits and motions

世界模型

[Orbis 2] Orbis 2: A Hierarchical World Model for Driving

[en] Causally Debiased Latent Action Model for Embodied Action Conditioned World Models

[PhysMani] PhysMani: Physics-principled 3D World Model for Dynamic Object Manipulation

[ACID] ACID: Action Consistency via Inverse Dynamics for Planning with World Models

Sim-to-Real

[Learning] Actuator Reality Shaping for Zero-Shot Sim-to-Real Robot Learning

[en] Efficiently Linking Real Scenes with Synthetic Data Generation for AI-based Cognitive Robotics and Computer Vision Applications

Sim-to-Real Transfer for Muscle-Actuated Robots via Generalized Actuator Networks

Can Context Bridge the Reality Gap? Sim-to-Real Transfer of Context-Aware Policies

基础理论

Robustifying Vision-Language Models via Test-Time Prompt Adaptation

[HPSv3++] HPSv3++: Scaling Reward Models Across the Full Spectrum of Diffusion Model Capabilities

When Robots Should Say "I Don't Know": Benchmarking Abstention in Embodied Question Answering

[en] RoboChallenge: Large-scale Real-robot Evaluation of Embodied Policies