Safe Reinforcement Learning for a Robot Being Pursued but with Objectives Covering More Than Capture-avoidance

TL;DR

This paper develops a safe reinforcement learning framework for robots in pursuit scenarios, ensuring safety beyond simple capture-avoidance, with validation through simulations and experiments.

Contribution

It introduces a novel safe RL framework that provides safety guarantees for pursuit scenarios involving multiple objectives, extending beyond traditional capture-avoidance.

Findings

The framework effectively ensures safety in pursuit scenarios.

Simulations and experiments validate the approach's robustness.

The method addresses safety concerns in real-world robotic applications.

Abstract

Reinforcement Learning (RL) algorithms show amazing performance in recent years, but placing RL in real-world applications such as self-driven vehicles may suffer safety problems. A self-driven vehicle moving to a target position following a learned policy may suffer a vehicle with unpredictable aggressive behaviors or even being pursued by a vehicle following a Nash strategy. To address the safety issue of the self-driven vehicle in this scenario, this paper conducts a preliminary study based on a system of robots. A safe RL framework with safety guarantees is developed for a robot being pursued but with objectives covering more than capture-avoidance. Simulations and experiments are conducted based on the system of robots to evaluate the effectiveness of the developed safe RL framework.