惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

IT之家
IT之家
The GitHub Blog
The GitHub Blog
F
Fortinet All Blogs
Last Week in AI
Last Week in AI
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
L
LangChain Blog
爱范儿
爱范儿
博客园_首页
Stack Overflow Blog
Stack Overflow Blog
MongoDB | Blog
MongoDB | Blog
博客园 - 三生石上(FineUI控件)
大猫的无限游戏
大猫的无限游戏
宝玉的分享
宝玉的分享
GbyAI
GbyAI
H
Help Net Security
A
About on SuperTechFans
Recent Announcements
Recent Announcements
Hugging Face - Blog
Hugging Face - Blog
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
雷峰网
雷峰网
D
Docker
博客园 - Franky
有赞技术团队
有赞技术团队
G
Google Developers Blog

cs.RO updates on arXiv.org

FineCog-Nav: Integrating Fine-grained Cognitive Modules for Zero-shot Multimodal UAV Navigation DENALI: A Dataset Enabling Non-Line-of-Sight Spatial Reasoning with Low-Cost LiDARs SENSE: Stereo OpEN Vocabulary SEmantic Segmentation Continual Hand-Eye Calibration for Open-world Robotic Manipulation PLAF: Pixel-wise Language-Aligned Feature Extraction for Efficient 3D Scene Understanding GaussianFlow SLAM: Monocular Gaussian Splatting SLAM Guided by GaussianFlow GIST: Multimodal Knowledge Extraction and Spatial Grounding via Intelligent Semantic Topology $π_{0.7}$: a Steerable Generalist Robotic Foundation Model with Emergent Capabilities R3D: Revisiting 3D Policy Learning Vision-Based Safe Human-Robot Collaboration with Uncertainty Guarantees Benchmarking Classical Coverage Path Planning Heuristics on Irregular Hexagonal Grids for Maritime Coverage Scenarios NEAT-NC: NEAT guided Navigation Cells for Robot Path Planning HRDexDB: A Large-Scale Dataset of Dexterous Human and Robotic Hand Grasps ADAPT: Benchmarking Commonsense Planning under Unspecified Affordance Constraints An Intelligent Robotic and Bio-Digestor Framework for Smart Waste Management Efficient closed-form approaches for pose estimation using Sylvester forms World-Value-Action Model: Implicit Planning for Vision-Language-Action Systems A Nonasymptotic Theory of Gain-Dependent Error Dynamics in Behavior Cloning CooperDrive: Enhancing Driving Decisions Through Cooperative Perception SpaceMind: A Modular and Self-Evolving Embodied Vision-Language Agent Framework for Autonomous On-orbit Servicing HiVLA: A Visual-Grounded-Centric Hierarchical Embodied Manipulation System UMI-3D: Extending Universal Manipulation Interface from Vision-Limited to 3D Spatial Perception Towards Multi-Object-Tracking with Radar on a Fast Moving Vehicle: On the Potential of Processing Radar in the Frequency Domain Beyond Conservative Automated Driving in Multi-Agent Scenarios via Coupled Model Predictive Control and Deep Reinforcement Learning Failure Identification in Imitation Learning Via Statistical and Semantic Filtering A Dynamic-Growing Fuzzy-Neuro Controller, Application to a 3PSP Parallel Robot Vision-Language-Action Jump-Starting for Reinforcement Learning Robotic Agents A Mechanistic Analysis of Sim-and-Real Co-Training in Generative Robot Policies ESCAPE: Episodic Spatial Memory and Adaptive Execution Policy for Long-Horizon Mobile Manipulation Evolvable Embodied Agent for Robotic Manipulation via Long Short-Term Reflection and Optimization
Off-Policy Safe Reinforcement Learning with Constrained O...
Guopeng Li, Matthijs T. J. Spaan, Julian F. P. Kooij · 2026-03-25 · via cs.RO updates on arXiv.org

When safety is formulated as a limit of cumulative cost, safe reinforcement learning (RL) aims to learn policies that maximize return subject to the cost constraint in data collection and deployment. Off-policy safe RL methods, although offering high sample efficiency, suffer from constraint violations due to cost-agnostic exploration and estimation bias in cumulative cost. To address this issue, we propose Constrained Optimistic eXploration Q-learning (COX-Q), an off-policy safe RL algorithm that integrates cost-bounded online exploration and conservative offline distributional value learning. First, we introduce a novel cost-constrained optimistic exploration strategy that resolves gradient conflicts between reward and cost in the action space and adaptively adjusts the trust region to control the training cost. Second, we adopt truncated quantile critics to stabilize the cost value learning. Quantile critics also quantify epistemic uncertainty to guide exploration. Experiments on safe velocity, safe navigation, and autonomous driving tasks demonstrate that COX-Q achieves high sample efficiency, competitive test safety performance, and controlled data collection cost. The results highlight COX-Q as a promising RL method for safety-critical applications.