Robot Policy Generalization

Follow this topic Ask about this topic

Robot policy generalization asks whether a controller can act reliably when objects, scenes, instructions, camera views, robot bodies, or dynamics differ from training. Approaches include collecting diverse demonstrations and using simulation randomization, structured object- or geometry-centered representations, modular skills, and vision-language-action or world-model policies. Researchers also use reinforcement-learning fine-tuning, test-time adaptation, and cross-embodiment action representations to help policies recover from unfamiliar conditions. A central challenge is that strong performance on familiar tasks or isolated shifts does not guarantee compositional generalization, so benchmarks increasingly test multiple interacting changes and real-world transfer.

Latest papers

179 papers20 papers in the last four weeks, up 150% on the four weeks before. 0.1% of all new papers.

Beyond Imitation: Reinforcement Learning Fine-Tuning for Adaptive Diffusion Navigation Policies
Mar 13, 2026Junhe Sheng, Ruofei Bai, Kuan Xu +5Robot Navigation Robot Policy Adaptation
Nanyang Technological University, Singapore. · A*STAR, Singapore. · National University of Singapore, Singapore.
Grounding Robot Generalization in Training Data via Retrieval-Augmented VLMs
Mar 12, 2026Jensen Gao, Dorsa Sadigh, Sandy Huang +1Robotic Policy Evaluation Robot Policy Generalization
Stanford University · Google DeepMind · Princeton University
APPLV: Adaptive Planner Parameter Learning from Vision-Language-Action Model
Mar 9, 2026Yuanjie Lu, Beichen Wang, Zhengqi Wu +4Robot Navigation Robot Policy Generalization
Department of Computer Science, George Mason University, Virginia, USA · Department of Engineering Science, University of South Florida, Florida, USA · Department of Computer Science, Rutgers University, New Jersey, USA
Seed2Scale: A Self-Evolving Data Engine with Parallel Worlds Expansion for Scalable Robot Learning
Mar 9, 2026Cong Tai, Zhaoyu Zheng, Haixu Long +12Synthetic Data Generation for Robotics Data-Efficient Robot Learning
ZTE Corporation, China
RoboRouter: Training-Free Policy Routing for Robotic Manipulation
Mar 9, 2026Yiteng Chen, Zhe Cao, Hongjia Ren +9Robot Policy Learning Robotic Manipulation
South China University of Technology · Nanjing University · University of New South Wales +5
Foundation and Small Models Coordination for Visuomotor Policy Learning
Mar 9, 2026Haoran Ding, Liang Ma, Yaxun Yang +7Visuomotor Policy Learning Robot Policy Generalization
Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), Abu Dhabi, UAE
Don't Freeze, Don't Crash: Extending the Safe Operating Range of Neural Navigation in Dense Crowds
Mar 6, 2026Jiefu Zhang, Yang Xu, Vaneet AggarwalSafe Robot Navigation Robot Navigation
Purdue University, USA
Task Parameter Extrapolation via Learning Inverse Tasks from Forward Demonstrations
Mar 5, 2026Serdar Bahar, Fatih Dogangun, Matteo Saveriano +2Robot Skill Learning Generalization in Robotic Manipulation
Department of Computer Engineering, Bogazici University, Istanbul, Türkiye. · Department of Industrial Engineering, University of Trento, Trento, Italy. · IRCN, the University of Tokyo, Tokyo, Japan.
From Local Corrections to Generalized Skills: Improving Neuro-Symbolic Policies with MEMO
Mar 4, 2026Benjamin A. Christie, Yinlong Dai, Mohammad Bararjanianbahnamiri +2Robot Skill Learning Language-Conditioned Robot Manipulation
Collab, Dept. of Mechanical Engineering, Virginia Tech, Blacksburg, USA. · TEA Lab, Dept. of Mechanical Engineering, Virginia Tech, Blacksburg, USA.
HALO: Learning Human-Robot Collaboration via Heterogeneous-Agent Lyapunov Policy Optimization
Mar 4, 2026Hao Zhang, Yaru Niu, Yikai Wang +2Decentralized MARL Heterogeneous MARL
ETAIC Lab, Department of Electrical Engineering, University of Texas at Arlington, USA · Safe AI Lab, Department of Mechanical Engineering, Carnegie Mellon University, USA.
Rethinking Camera Choice: An Empirical Study on Fisheye Camera Properties in Robotic Manipulation
Mar 2, 2026Han Xue, Nan Min, Xiaotong Liu +5Robot Policy Generalization Robot Imitation Learning
Shanghai Jiao Tong University · Southeast University · USTC +2
LessMimic: Long-Horizon Humanoid Interaction with Unified Distance Field Representations
Feb 25, 2026Yutang Lin, Jieming Cui, Yixuan Li +3Humanoid Robot Control Long-Horizon Robotic Manipulation
Generalizing from References using a Multi-Task Reference and Goal-Driven RL Framework
Feb 23, 2026Jiashun Wang, M. Eva Mungai, He Li +3Humanoid Robot Control Robot Policy Generalization
RAI Institute · Carnegie Mellon University
Universal Pose Pretraining for Generalizable Vision-Language-Action Policies
Feb 23, 2026Haitao Lin, Hanyang Yu, Jingshun Huang +5Robot Policy Learning Vision-Language-Action Models
Tencent Robotics X · Futian Laboratory · The Hong Kong University of Science and Technology +2
FRAPPE: Infusing World Modeling into Generalist Policies via Multiple Future Representation Alignment
Feb 19, 2026Han Zhao, Jingbo Wang, Wenxuan Song +5World Models for Robotics Latent World Models
Zhejiang University · Westlake University · HKUST (GZ) +3
SimToolReal: An Object-Centric Policy for Zero-Shot Dexterous Tool Manipulation
Feb 18, 2026Kushal Kedia, Tyler Ga Wei Lum, Jeannette Bohg +1Zero-Shot Robotic Manipulation Robot Skill Learning
Cornell University · Stanford University
BPP: Long-Context Robot Imitation Learning by Focusing on Key History Frames
Feb 16, 2026Max Sobol Mark, Jacky Liang, Maria Attarian +4Robot Imitation Learning Robot Learning
Carnegie Mellon University · Google DeepMind · Princeton University
Learning Transferability: A Two-Stage Reinforcement Learning Approach for Enhancing Quadruped Robots' Performance in U-Shaped Stair Climbing
Feb 16, 2026Baixiao Huang, Baiyu Huang, Yu HouRL for Legged Locomotion Quadruped Locomotion
Independent Researcher, 7200 Bollinger Rd, Apt 516, CA 95129. · Independent Researcher, 16506 82nd Pl NE, Kenmore, WA, 98028. · Western New England University, Department of Construction Management, 1215 Wilbraham Rd, Springfield, MA 01119 (corresponding author).
ReaDy-Go: Real-to-Sim Dynamic 3D Gaussian Splatting Simulation for Environment-Specific Visual Navigation with Moving Obstacles
Feb 12, 2026Seungyeon Yoo, Youngseok Jang, Dabin Kim +3Robotics Simulation Robot Navigation
Department of Aerospace Engineering, Seoul National University, Seoul, Republic of Korea · InnoCORE AI-Transformed Aerospace Research Center, KAIST, Daejeon, Republic of Korea
Spotlighting Task-Relevant Features: Object-Centric Representations for Better Generalization in Robotic Manipulation
Jan 29, 2026Alexandre Chapin, Bruno Machado, Emmanuel Dellandréa +1Visual Representation Learning Generalization in Robotic Manipulation
LIRIS, imagine
Agile and Generalized Legged Locomotion via Attention-Based Neural Map Encoding
Jan 13, 2026Chong Zhang, Victor Klemm, Fan Yang +1Terrain-Aware Robot Locomotion RL for Legged Locomotion
1Robotic Systems Lab, ETH Zurich, Switzerland · 2Secure, Reliable, and Intelligent Systems Lab, ETH Zurich, Switzerland · 3ETH AI Center, Switzerland
Beyond Topology: A Morphological Symmetry Graph Representation for Locomotion Policy Learning
Nov 30, 2025Sizhe Wei, Xulin Chen, Fengze Xie +3RL for Legged Locomotion Equivariant GNNs
Georgia Institute of Technology · Syracuse University · California Institute of Technology
Sampling Strategies for Robust Universal Quadrupedal Locomotion Policies
Oct 8, 2025David Rytz, Kim Tien Ly, Ioannis HavoutisDomain Randomization Sim-to-Real Transfer
1Dynamic Robot Systems, Oxford Robotics Institute, University of Oxford
MM-Nav: Multi-View VLA Model for Robust Visual Navigation via Multi-Expert Learning
Oct 3, 2025Tianyu Xu, Jiawei Chen, Jiazhao Zhang +5Robot Navigation Visual Robot Navigation
Peking University, Beijing, China · Galbot, Beijing, China · Shanghai Jiao Tong University, Shanghai, China +1
Agentic Scene Policies
Sep 23, 2025Sacha Morin, Kumaraditya Gupta, Mahtab Sandhu +4Robot Policy Generalization Robot Learning
Université de Montréal · Mila - Quebec AI Institute · Sapienza University of Rome
Rational Inverse Reasoning: Few-Shot Imitation by Inferring Intent through Planning
Aug 12, 2025Ben Zandonati, Tomás Lozano-Pérez, Leslie Pack KaelblingRobot Skill Learning Generalization in Robotic Manipulation
MIT CSAIL
Shared Modular Recurrence in Contextual MDPs for Universal Morphology Control
Jun 10, 2025Laurens Engwegen, Max Weltevrede, Caroline Horsch +2Robotic Control Reinforcement Learning
Delft University of Technology Delft, The Netherlands
An Real-Sim-Real (RSR) Loop Framework for Generalizable Robotic Policy Transfer
Mar 13, 2025Yuxuan Xu, Shiyu Wang, Jinhao Huang +6Robotic Data Collection Sim-to-Real Transfer
Institute for AI Industry Research (AIR), Tsinghua University · Beijing Jiaotong University · The Hong Kong University of Science and Technology (Guangzhou)
Data Scaling Laws in Imitation Learning for Robotic Manipulation
Oct 24, 2024Fanqi Lin, Yingdong Hu, Pingyue Sheng +3Generalization in Robotic Manipulation Robotic Manipulation
Tsinghua University · Shanghai Qi Zhi Institute · Shanghai Artificial Intelligence Laboratory

Mar 13, 2026J/K move · Enter open · S save

Beyond Imitation: Reinforcement Learning Fine-Tuning for Adaptive Diffusion Navigation Policies

Junhe Sheng, Ruofei Bai, Kuan Xu, Ruimeng Liu+4

Nanyang Technological University, Singapore. · A*STAR, Singapore. · National University of Singapore, Singapore.

Robot Navigation Robot Policy Adaptation RL Fine-Tuning Diffusion Policy Robot Policy Generalization

Open paper AskPDF arXiv

Diffusion-based robot navigation policies trained on large-scale imitation learning datasets, can generate multi-modal trajectories directly from the robot's visual observations, bypassing the traditional localization-mapping-planning pipeline and achieving strong zero-shot generalization. However, their performance remains constrained by the coverage of offline datasets, and when deployed in unseen settings, distribution shift often leads to accumulated trajectory errors and safety-critical failures. Adapting diffusion policies with reinforcement learning is challenging because their iterative denoising structure hinders effective gradient backpropagation, while also making the training of an additional value network computationally expensive and less stable. To address these issues, we propose a reinforcement learning fine-tuning framework tailored for diffusion-based navigation. The method leverages the inherent multi-trajectory sampling mechanism of diffusion models and adopts Group Relative Policy Optimization (GRPO), which estimates relative advantages across sampled trajectories without requiring a separate value network. To preserve pretrained representations while enabling adaptation, we freeze the visual encoder and selectively update the higher decoder layers and action head, enhancing safety-aware behaviors through online environmental feedback. On the PointGoal task in Isaac Sim, our approach improves the Success Rate from 52.0% to 58.7% and SPL from 0.49 to 0.54 on unseen scenes, while reducing collision frequency. Additional experiments show that the fine-tuned policy transfers zero-shot to a real quadruped platform and maintains stable performance in geometrically out-of-distribution environments, suggesting improved adaptability and safe generalization to new domains.