Simple Models, Real Swimming: Digital Twins for Tendon-Driven Underwater Robots
Mike Yan Michelis, Nana Obayashi, Josie Hughes, Robert Kevin Katzschmann
AI summary
Problem
Modeling soft underwater robots is hindered by complex fluid-structure interactions and computationally expensive simulations that prevent scalable control and learning.
Approach
We implement a simplified, stateless hydrodynamics model in MuJoCo and calibrate its five fluid parameters using only two real-world swimming trajectories to create a fast digital twin.
Key results
- Accurate sim-to-real matching across actuation frequencies
- Outperforms classical Elongated Body Theory baseline
- 15x real-time simulation speed enabling parallel scaling
- 93% success rate in reinforcement learning target tracking
Why it matters
Offers an open-source, computationally efficient digital twin that lowers the barrier for scalable reinforcement learning and control in soft underwater robotics.
Abstract
Mimicking the graceful motion of swimming an- imals remains a core challenge in soft robotics due to the complexity of fluid-structure interaction and the difficulty of controlling soft, biomimetic bodies. Existing modeling ap- proaches are often computationally expensive and impractical for complex control or reinforcement learning needed for realistic motions to emerge in robotic systems. In this work, we present a tendon-driven fish robot modeled in an efficient underwater swimmer environment using a simplified, stateless hydrodynamics formulation implemented in the widespread robotics framework MuJoCo. With just two real-world swim- ming trajectories, we identify five fluid parameters that allow a matching to experimental behavior and generalize across a range of actuation frequencies. We show that this stateless fluid model can generalize to unseen actuation and outperform classical analytical models such as the elongated body theory. This simulation environment runs faster than real-time and can easily enable downstream learning algorithms such as reinforcement learning for target tracking, reaching a 93% success rate. Due to the simplicity and ease of use of the model and our open-source simulation environment, our results show that even simple, stateless models — when carefully matched to physical data — can serve as effective digital twins for soft underwater robots, opening up new directions for scalable learning and control in aquatic environments.