Egocentric Station Holding of Robotic Fish in Unknown Turbulent Background Flow

2026-07-26

AI Summary

This study addresses the challenge of precise station-holding for bio-inspired robotic fish in unknown turbulent flows, where strong nonlinear fluid–structure interactions severely hinder control performance. To overcome this limitation, the authors propose the SWiFT framework, which integrates free-swimming experiments, high-fidelity computational fluid dynamics (CFD) simulations, and a systematic sim-to-real transfer pipeline to develop a proprioception-only station-holding control policy via reinforcement learning. Notably, this approach is the first to replicate biological rheotaxis without explicit flow sensing, achieving stable station-holding in turbulence and substantially outperforming existing methods—evidenced by a significant reduction in root-mean-square error (RMSE) of positional deviation. The work establishes a scalable foundation for intelligent deployment of underwater robots in real-world, complex aquatic environments.

Abstract

Approaching a target position and holding station in flowing water is a fundamental and critical capability for robotic fish operating in natural aquatic environments. Despite decades of advances in enhancing swimming efficiency and maneuverability, this capability remains underdeveloped, largely owing to the insufficiently characterized, highly nonlinear fluid-structure interactions inherent to freely swimming robotic fish in flows. To bridge this gap, we propose the SWiFT framework, a Swimming With Flow Toolbox that enables the efficient exploration of an egocentric station-holding policy for a body and/or caudal fin (BCF) robotic fish in unknown and turbulent background flows via reinforcement learning (RL). Our SWiFT integrates a free-swimming flow-tank experimental setup with a highly efficient, physically consistent computational fluid dynamics (CFD)-based simulator and a systematic sim-to-real transfer pipeline. The resulting policy achieves substantial improvements over state-of-the-art methods across all metrics, most notably root-mean-square error (RMSE) of distance. Furthermore, we validated that egocentric feedback alone, without any explicit flow sensing, enables station-holding in unknown turbulent flows, closely mirroring the biological phenomenon of rheotaxis. Accordingly, the success of this egocentric station-holding policy not only advances robotic fish control toward real-world deployment, but also highlights SWiFT's promise as a foundation for tackling complex swimming tasks for underwater robots.

Problem

• egocentric station holding
• robotic fish
• turbulent flow
• rheotaxis
• fluid-structure interaction

Innovation

• egocentric station-holding
• reinforcement learning
• computational fluid dynamics (CFD)
• sim-to-real transfer
• rheotaxis