Learning Flow-Adaptive Dynamic Model for Robotic Fish Swimming in Unknown Background Flow

Kaitian Chao; Xiaozhu Lin; Xiaopei Liu; Yang Wang

IROS 20250 citations

Learning Flow-Adaptive Dynamic Model for Robotic Fish Swimming in Unknown Background Flow

Kaitian Chao, Xiaozhu Lin, Xiaopei Liu, Yang Wang

Abstract

Robotic fish face considerable challenges in natural environment due to the absence of a comprehensive and precise model that can depict the intricate fluid-structure interactions, particularly in the presence of background flow fields. To this end, we present a novel data-driven dynamic modeling framework capable of characterizing the swimming motions of the robotic fish under various background flow conditions without the necessity for explicit flow information. The model is synthesized by an internal model with an adaptive residual acceleration model to effectively isolate and address external flow effects. Notably, the residual model employs the innovative Domain Adversarially Invariant Meta-Learning (DAIML) approach, allowing the framework to adapt to fluctuating and previously unseen background flow scenarios, enhancing its robustness and scalability. Validation through high-fidelity Computational Fluid Dynamics (CFD) simulations demonstrates the framework’s effectiveness in improving the performance of robotic fish across diverse real-world aquatic environments.

BibTeX

@inproceedings{iros2025_learningflowadap,
  title = {Learning Flow-Adaptive Dynamic Model for Robotic Fish Swimming in Unknown Background Flow},
  author = {Kaitian Chao and Xiaozhu Lin and Xiaopei Liu and Yang Wang},
  booktitle = {IROS 2025},
  year = {2025}
}