A critical assessment of reinforcement learning methods for microswimmer navigation in complex flows

TL;DR

This paper critically evaluates reinforcement learning algorithms for microswimmer navigation in complex fluid flows, demonstrating that advanced methods like PPO outperform traditional algorithms and can achieve near-optimal performance with proper tuning.

Contribution

It provides a comprehensive assessment of RL methods in fluid navigation, highlighting the superiority of PPO and emphasizing the importance of implementation details and hyperparameter tuning.

Findings

PPO outperforms Q-Learning and Advantage Actor Critic in complex flows.

Proper tuning and techniques enable PPO to match theoretical optimal performance.

Traditional RL algorithms show poor robustness in turbulent flow environments.

Abstract

Navigating in a fluid flow while being carried by it, using only information accessible from on-board sensors, is a problem commonly faced by small planktonic organisms. It is also directly relevant to autonomous robots deployed in the oceans. In the last ten years, the fluid mechanics community has widely adopted reinforcement learning, often in the form of its simplest implementations, to address this challenge. But it is unclear how good are the strategies learned by these algorithms. In this paper, we perform a quantitative assessment of reinforcement learning methods applied to navigation in partially observable flows. We first introduce a well-posed problem of directional navigation for which a quasi-optimal policy is known analytically. We then report on the poor performance and robustness of commonly used algorithms (Q-Learning, Advantage Actor Critic) in flows regularly encountered in the literature: Taylor-Green vortices, Arnold-Beltrami-Childress flow, and two-dimensional turbulence. We show that they are vastly surpassed by PPO (Proximal Policy Optimization), a more advanced algorithm that has established dominance across a wide range of benchmarks in the reinforcement learning community. In particular, our custom implementation of PPO matches the theoretical quasi-optimal performance in turbulent flow and does so in a robust manner. Reaching this result required the use of several additional techniques, such as vectorized environments and generalized advantage estimation, as well as hyperparameter optimization. This study demonstrates the importance of algorithm selection, implementation details, and fine-tuning for discovering truly smart autonomous navigation strategies in complex flows.