Safe Reinforcement Learning with Hard Constraints: Policy Gradient Methods with Intersample Guarantees
Source: arxiv.org/abs/2611.21987
We develop a policy-gradient (PG) reinforcement learning (RL) framework with hard inter-sample control barrier function (CBF)-based safety constraints for nonlinear sampled-data systems. We first design a class of stochastic policies with support truncated to sampling-period-dependent safe control sets, that guarantee safety-by-construction at sampling instants. We then propose a model-free PG algorithm that searches over the class of safe-by-design policies and establish sufficient conditions on sampling-period such that any action sampled from safe control sets remains safe throughout the…
Comments and reviews
No comments yet.
API
curl /api/v1/records/arxiv/2611.21987
curl -X POST /api/v1/records/arxiv/2611.21987/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2611.21987/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"body": "…", "review": true}'