Guided Safe Shooting: model based reinforcement learning with safety constraints

TL;DR

Guided Safe Shooting (GuSS) is a model-based reinforcement learning method that ensures safety constraints are minimally violated while efficiently exploring and learning control policies for real-world systems.

Contribution

The paper introduces GuSS, a novel model-based RL approach with safety guarantees and three safe planning algorithms, improving safety and efficiency over existing methods.

Findings

GuSS reduces unsafe states during learning.

GuSS achieves high rewards with fewer real-system interactions.

Safe planners enable effective exploration and safety compliance.

Abstract

In the last decade, reinforcement learning successfully solved complex control tasks and decision-making problems, like the Go board game. Yet, there are few success stories when it comes to deploying those algorithms to real-world scenarios. One of the reasons is the lack of guarantees when dealing with and avoiding unsafe states, a fundamental requirement in critical control engineering systems. In this paper, we introduce Guided Safe Shooting (GuSS), a model-based RL approach that can learn to control systems with minimal violations of the safety constraints. The model is learned on the data collected during the operation of the system in an iterated batch fashion, and is then used to plan for the best action to perform at each time step. We propose three different safe planners, one based on a simple random shooting strategy and two based on MAP-Elites, a more advanced divergent-search algorithm. Experiments show that these planners help the learning agent avoid unsafe situations while maximally exploring the state space, a necessary aspect when learning an accurate model of the system. Furthermore, compared to model-free approaches, learning a model allows GuSS reducing the number of interactions with the real-system while still reaching high rewards, a fundamental requirement when handling engineering systems.