mlsanityData to 14 Oct 2026

Learning over Forward-Invariant Policy Classes: Reinforcement Learning without Safety Concerns

arXiv 2604.07875 ↗First posted 9 Apr 2026eess.SY cs.SYcited by 0

On the map: Deep reinforcement learning methods → Safe and constrained RL

Abstract (arXiv)

This paper proposes a safe reinforcement learning (RL) framework based on forward-invariance-induced action-space design. The control problem is cast as a Markov decision process, but instead of relying on runtime shielding or penalty-based constraints, safety is embedded directly into the action representation. Specifically, we construct a finite admissible action set in which each discrete action corresponds to a stabilizing feedback law that preserves forward invariance of a prescribed safe state set. Consequently, the RL agent optimizes policies over a safe-by-construction policy class. We validate the framework on a quadcopter hover-regulation problem under disturbance. Simulation results show that the learned policy improves closed-loop performance and switching efficiency, while all evaluated policies remain safety-preserving. The proposed formulation decouples safety assurance from performance optimization and provides a promising foundation for safe learning in nonlinear systems.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Safe and constrained RL, one of 15 directions in Deep reinforcement learning methods. Builds on 8 of its 24 references that are in the map. First posted on arXiv 9 Apr 2026.

Builds on

8 of 24 references.

References · 8

  1. 2017Constrained Policy Optimizationcited by 983
  2. 2017Safe Model-based Reinforcement Learning with Stability Guaranteescited by 366
  3. 2018A Lyapunov-based Approach to Safe Reinforcement Learningcited by 297
  4. 2019End-to-End Safe Reinforcement Learning through Barrier Functions for Safety-Critical Continuous Control Taskscited by 289
  5. 2016Control Barrier Function Based Quadratic Programs for Safety Critical Systemscited by 287
  6. 2018Reinforcement Learning for UAV Attitude Controlcited by 48
  7. 2016Safe Learning of Regions of Attraction for Uncertain, Nonlinear Systems with Gaussian Processescited by 39
  8. 2017Differential Flatness of Quadrotor Dynamics Subject to Rotor Drag for Accurate Tracking of High-Speed Trajectoriescited by 32

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024Sampling-based Safe Reinforcement Learning for Nonlinear Dynamical Systems
  2. 2026Safety Guarantees in Zero-Shot Reinforcement Learning for Cascade Dynamical Systems
  3. 2026SafePG: Safe and Globally Optimal Reinforcement Learning with Hard Constraints
  4. 2026Towards Safe Reinforcement Learning with Reduced Conservativeness: A Case Study on Drone Flight Control
  5. 2024Safety Filtering While Training: Improving the Performance and Sample Efficiency of Reinforcement Learning Agents
  6. 2024Safe Reinforcement Learning Filter for Multicopter Collision-Free Tracking under disturbances
  7. 2026Learning When to Act: Communication-Efficient Reinforcement Learning via Run-Time Assurance
  8. 2026Safe Reinforcement Learning using Ideas from Model Predictive Control