Certificated Actor-Critic: Hierarchical Reinforcement Learning with Control Barrier Functions for Safe Navigation

TL;DR

This paper introduces a novel model-free hierarchical reinforcement learning algorithm called Certificated Actor-Critic (CAC) that integrates Control Barrier Functions for safe robot navigation, addressing limitations of existing CBF methods.

Contribution

The paper proposes a new CAC algorithm combining hierarchical RL and CBF-derived rewards, with theoretical validation and implementation improvements.

Findings

CAC effectively ensures safety in simulated navigation tasks.

Theoretical analysis confirms convergence and safety guarantees.

Simulation results demonstrate improved navigation performance.

Abstract

Control Barrier Functions (CBFs) have emerged as a prominent approach to designing safe navigation systems of robots. Despite their popularity, current CBF-based methods exhibit some limitations: optimization-based safe control techniques tend to be either myopic or computationally intensive, and they rely on simplified system models; conversely, the learning-based methods suffer from the lack of quantitative indication in terms of navigation performance and safety. In this paper, we present a new model-free reinforcement learning algorithm called Certificated Actor-Critic (CAC), which introduces a hierarchical reinforcement learning framework and well-defined reward functions derived from CBFs. We carry out theoretical analysis and proof of our algorithm, and propose several improvements in algorithm implementation. Our analysis is validated by two simulation experiments, showing the effectiveness of our proposed CAC algorithm.