mlsanityData to 7 Oct 2026

Safe Reinforcement Learning Using Black-Box Reachability Analysis

arXiv 2204.07417 ↗First posted 15 Apr 2022cs.RO cs.LG cs.SYcited by 22

Not on the map. Why?

Abstract (arXiv)

Reinforcement learning (RL) is capable of sophisticated motion planning and control for robots in uncertain environments. However, state-of-the-art deep RL approaches typically lack safety guarantees, especially when the robot and environment models are unknown. To justify widespread deployment, robots must respect safety constraints without sacrificing performance. Thus, we propose a Black-box Reachability-based Safety Layer (BRSL) with three main components: (1) data-driven reachability analysis for a black-box robot model, (2) a trajectory rollout planner that predicts future actions and observations using an ensemble of neural networks trained online, and (3) a differentiable polytope collision check between the reachable set and obstacles that enables correcting unsafe actions. In simulation, BRSL outperforms other state-of-the-art safe RL methods on a Turtlebot 3, a quadrotor, a trajectory-tracking point mass, and a hexarotor in wind with an unsafe set adjacent to the area of highest reward.

Key facts. Cited by 22 papers in the map (data through 7 Oct 2026). Builds on 33 of its 59 references that are in the map. First posted on arXiv 15 Apr 2022.

Builds on

33 of 59 references.

References · 33

  1. 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,635
  2. 2015Continuous control with deep reinforcement learningcited by 3,279
  3. 2017Constrained Policy Optimizationcited by 928
  4. 2018Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Modelscited by 748
  5. 2017OptNet: Differentiable Optimization as a Layer in Neural Networkscited by 588
  6. 2018Reward Constrained Policy Optimizationcited by 397
  7. 2015Risk-Constrained Reinforcement Learning with Percentile Risk Criteriacited by 320
  8. 2018A Lyapunov-based Approach to Safe Reinforcement Learningcited by 278
  9. 2019End-to-End Safe Reinforcement Learning through Barrier Functions for Safety-Critical Continuous Control Taskscited by 256
  10. 2018Safe Exploration in Continuous Action Spacescited by 237
  11. 2020Exploration-Exploitation in Constrained MDPscited by 146
  12. 2020Provably Efficient Safe Exploration via Primal-Dual Policy Optimizationcited by 131
21 more
  1. 2013Risk-sensitive Reinforcement Learningcited by 129
  2. 2017Socially Aware Motion Planning with Deep Reinforcement Learningcited by 128
  3. 2018Learning-based Model Predictive Control for Safe Explorationcited by 116
  4. 2020Conservative Safety Critics for Explorationcited by 107
  5. 2016Decentralized Non-communicating Multiagent Collision Avoidance with Deep Reinforcement Learningcited by 104
  6. 2017Virtual-to-real Deep Reinforcement Learning: Continuous Control of Mobile Robots for Mapless Navigationcited by 90
  7. 2017Towards Optimally Decentralized Multi-Robot Collision Avoidance via Deep Reinforcement Learningcited by 81
  8. 2017Socially Compliant Navigation through Raw Depth Inputs with Generative Adversarial Imitation Learningcited by 50
  9. 2021Safe Reinforcement Learning Using Advantage-Based Interventioncited by 45
  10. 2019Value constrained model-free continuous controlcited by 34
  11. 2020Reachability-based Trajectory Safeguard (RTS): A Safe and Fast Reinforcement Learning Safety Layer for Continuous Controlcited by 28
  12. 2020On Infusing Reachability-Based Safety Assurance within Planning Frameworks for Human-Robot Vehicle Interactionscited by 24
  13. 2020Safe Active Dynamics Learning and Control: A Sequential Exploration-Exploitation Frameworkcited by 19
  14. 2018Bridging the Gap Between Safety and Real-Time Performance in Receding-Horizon Trajectory Design for Mobile Robotscited by 11
  15. 2020Data-Driven Reachability Analysis Using Matrix Zonotopescited by 11
  16. 2021Model-Free Safety-Critical Control for Robotic Systemscited by 8
  17. 2021Constrained Feedforward Neural Network Training via Reachability Analysiscited by 5
  18. 2021Control Barrier Functions With Unmodeled Dynamics Using Integral Quadratic Constraintscited by 4
  19. 2018Safely Probabilistically Complete Real-Time Planning and Exploration in Unknown Environmentscited by 3
  20. 2020Set operations and order reductions for constrained zonotopescited by 3
  21. 2021Model Free Barrier Functions via Implicit Evading Maneuverscited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 3; 2023: 3; 2024: 7; 2025: 6; 2026: 3.

22 papers.

Sort:

By area:

  1. 2023Iterative Reachability Estimation for Safe Reinforcement Learningcited by 33
  2. 2022Provably Safe Reinforcement Learning: Conceptual Analysis, Survey, and Benchmarkingcited by 31
  3. 2022Provably Safe Reinforcement Learning via Action Projection using Reachability Analysis and Polynomial Zonotopescited by 23
  4. 2024Safe Reinforcement Learning on the Constraint Manifold: Theory and Applicationscited by 12
  5. 2024Dynamic Model Predictive Shielding for Provably Safe Reinforcement Learningcited by 11
  6. 2024SoNIC: Safe Social Navigation with Adaptive Conformal Inference and Constrained Reinforcement Learningcited by 8
  7. 2025Safe LLM-Controlled Robots with Formal Guarantees via Reachability Analysiscited by 7
  8. 2024RAIL: Reachability-Aided Imitation Learning for Safe Policy Executioncited by 5
  9. 2022Safe Reinforcement Learning using Data-Driven Predictive Controlcited by 4
  10. 2024Bridging the gap between Learning-to-plan, Motion Primitives and Safe Reinforcement Learningcited by 4
  11. 2024Do No Harm: A Counterfactual Approach to Safe Reinforcement Learningcited by 1
  12. 2024AI-Powered CPS-Enabled Vulnerable-User-Aware Urban Transportation Digital Twin: Methods and Applicationscited by 1
10 more
  1. 2025Leveraging Analytic Gradients in Provably Safe Reinforcement Learningcited by 1
  2. 2025Predictive Safety Shield for Dyna-Q Reinforcement Learningcited by 1
  3. 2023Safe Deep RL for Intraoperative Planning of Pedicle Screw Placementcited by 0
  4. 2023Adaptive Policy Learning to Additional Taskscited by 0
  5. 2025Safe Learning Control with Optimality and Stability Guaranteescited by 0
  6. 2025Safely Learning Controlled Stochastic Dynamicscited by 0
  7. 2025Constraint-Aware Reinforcement Learning via Adaptive Action Scalingcited by 0
  8. 2026Conformal Reachability for Safe Control in Unknown Environmentscited by 0
  9. 2026Safe Continual Reinforcement Learning in Non-stationary Environmentscited by 0
  10. 2026No Free Checker: A Survey of Verifiers for Robot Policiescited by 0