Safe Reinforcement Learning with Receding-Horizon Safety Critic and Differentiable Action Shields

arXiv:2610.20255 · cs.RO · Submitted 2026-10-14 · Read on arXiv

cs.RO

Submitted: 2026-10-14

Updated: 2026-10-14

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Reinforcement Learning — 1× in this paper · explained in 4 episodes

Sources

• Differentiable Convex Optimization Layers
• Neural Control Barrier Functions from Physics Informed Neural Networks
• End-to-End Safe Reinforcement Learning through Barrier Functions for Safety-Critical Continuous Control Tasks
• Reinforcement Learning for Safe Robot Control using Control Lyapunov Barrier Functions
• Data-Driven Robust Barrier Functions for Safe, Long-Term Operation
• Safe Reinforcement Learning Using Robust Control Barrier Functions
• Multi-Layered Safety for Legged Robots via Control Barrier Functions and Model Predictive Control
• T-CBF: Traversability-based Control Barrier Function to Navigate Vertically Challenging Terrain
• Handling Long-Term Safety and Uncertainty in Safe Reinforcement Learning
• Agile But Safe: Learning Collision-Free High-Speed Legged Locomotion
• Safe Reinforcement Learning on the Constraint Manifold: Theory and Applications
• Safe Control Under Input Limits with Neural Control Barrier Functions
• Learning Barrier Functions with Memory for Robust Safe Navigation
• Reinforcement Learning-Enhanced Control Barrier Functions for Robot Manipulators
• SafeFall: Learning Protective Control for Humanoid Robots
• Learning Control Barrier Functions from Expert Demonstrations
• Neural Gaits: Learning Bipedal Locomotion via Control Barrier Functions and Zero Dynamics Policies
• Reinforcement Learning-based Receding Horizon Control using Adaptive Control Barrier Functions for Safety-Critical Systems
• Proximal Policy Optimization Algorithms
• How to Train Your Neural Control Barrier Function: Learning Safety Filters for Complex Input-Constrained Systems

Related papers

• FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
• MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
• RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
• ARC: A Reasoning Recipe for Robot Foundation Models
• HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
• APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies