← Back
ICRA 2026
Experimental Validations of a Digital Twin Model for Underwater Tracked Vehicle
Jong-Boo Han, Sangjin Lee, Yeongjun Lee, Daegil Park, Tae-Kyeong Yeu
AI summary
A high-fidelity digital twin model accurately replicates the locomotion of a tracked underwater robot over complex seabed terrain, achieving 97% position-level agreement with physical experiments.
Problem
Creating accurate digital twins for underwater tracked robots is difficult due to complex track-ground contact, hydrodynamic forces, and nonlinear dynamics, leaving many existing models insufficiently validated for real-world use.
Approach
The authors developed a multibody dynamics-based digital twin that integrates track-ground interaction and hydrodynamic effects, then validated it against water-tank experiments using a reconstructed seabed terrain and vision-based motion tracking.
Key results
- 97% position-level agreement between simulation and experiment
- 85% velocity-level agreement confirmed via cross-correlation
- Accurate reproduction of X/Z position and pitch angle trends over reconstructed terrain
- Validated framework enables reliable simulation-based controller tuning
Why it matters
Provides a validated simulation tool that reduces the cost and risk of real-world underwater robotics experiments while accelerating controller development and performance evaluation.
Abstract
No abstract on file.