Learning control of underactuated double pendulum with Model-Based Reinforcement Learning

TL;DR

This paper presents a control strategy for an underactuated double pendulum using a recent model-based reinforcement learning algorithm called MC-PILCO, focusing on implementation details for a robotics competition.

Contribution

The paper applies MC-PILCO to control an underactuated double pendulum, demonstrating its effectiveness in a robotics competition setting.

Findings

Successful application of MC-PILCO to the double pendulum task

Insights into critical implementation aspects of MC-PILCO

Potential for improved control in underactuated systems

Abstract

This report describes our proposed solution for the second AI Olympics competition held at IROS 2024. Our solution is based on a recent Model-Based Reinforcement Learning algorithm named MC-PILCO. Besides briefly reviewing the algorithm, we discuss the most critical aspects of the MC-PILCO implementation in the tasks at hand.