Equivariant Reinforcement Learning for Quadrotor UAV

TL;DR

This paper introduces an equivariant reinforcement learning framework for quadrotor UAVs that leverages symmetry properties to reduce training complexity and improve sample efficiency, demonstrated with TD3 and SAC algorithms.

Contribution

The paper proposes a novel equivariant RL approach exploiting quadrotor dynamics symmetry, reducing state dimensions and enhancing training efficiency.

Findings

Significant reduction in training samples needed for effective learning.

Improved sample efficiency demonstrated with TD3 and SAC algorithms.

The equivariant framework enhances RL applicability in resource-limited scenarios.

Abstract

This paper presents an equivariant reinforcement learning framework for quadrotor unmanned aerial vehicles. Successful training of reinforcement learning often requires numerous interactions with the environments, which hinders its applicability especially when the available computational resources are limited, or when there is no reliable simulation model. We identified an equivariance property of the quadrotor dynamics such that the dimension of the state required in the training is reduced by one, thereby improving the sampling efficiency of reinforcement learning substantially. This is illustrated by numerical examples with popular reinforcement learning techniques of TD3 and SAC.