Velocity-History-Based Soft Actor-Critic Tackling IROS'24 Competition "AI Olympics with RealAIGym"

TL;DR

This paper introduces a novel reinforcement learning approach based on Soft Actor-Critic with a history-encoding context vector, successfully applied to control chaotic underactuated systems in the IROS'24 competition.

Contribution

It extends SAC with a CNN-based context vector to incorporate history, improving control of chaotic systems in real-world competitions.

Findings

Achieved high performance scores in IROS'24 competition

Demonstrated robustness on Pendubot and Acrobot tasks

Outperformed baseline control algorithms

Abstract

The ``AI Olympics with RealAIGym'' competition challenges participants to stabilize chaotic underactuated dynamical systems with advanced control algorithms. In this paper, we present a novel solution submitted to IROS'24 competition, which builds upon Soft Actor-Critic (SAC), a popular model-free entropy-regularized Reinforcement Learning (RL) algorithm. We add a `context' vector to the state, which encodes the immediate history via a Convolutional Neural Network (CNN) to counteract the unmodeled effects on the real system. Our method achieves high performance scores and competitive robustness scores on both tracks of the competition: Pendubot and Acrobot.