Uncertainty-Aware Adaptive Dynamics for Underwater Vehicle�Manipulator Robots
Edward Morgan, Nenyi Kweku Dadson, Corina Barbalata
AI summary
Problem
Accurate dynamic modeling for underwater vehicle-manipulator systems is hindered by time-varying hydrodynamics and strong subsystem coupling. Existing adaptive methods struggle to update parameters online while guaranteeing physical plausibility and quantifying uncertainty.
Approach
The method uses moving horizon estimation to stack vehicle and manipulator regressors over a finite window, solving a convex optimization problem that enforces physical consistency constraints on inertia, damping, and hydrostatics while tracking parameter evolution.
Key results
- Manipulator torque predictions achieve R² = 0.88–0.98 with slopes near unity
- Vehicle surge, heave, and roll dynamics reproduced with good fidelity under strong coupling
- Median solver time of ~0.023 s per update confirms online feasibility
- Near 100% confidence interval coverage with rapidly converging, physically plausible parameters
Why it matters
Provides a reliable, uncertainty-quantified modeling foundation for real-time control and digital twins in unpredictable underwater environments.
Abstract
Accurate and adaptive dynamic models are critical for underwater vehicle–manipulator systems where hydrody- namic effects induce time-varying parameters. This paper introduces a novel uncertainty-aware adaptive dynamics model framework that remains linear in lumped vehicle and ma- nipulator parameters, and embeds convex physical consistency constraints during online estimation. Moving horizon estimation is used to stack horizon regressors, enforce realizable inertia, damping, friction, and hydrostatics, and quantify uncertainty from parameter evolution. Experiments on a BlueROV2 Heavy with a 4-DOF manipulator demonstrate rapid convergence and calibrated predictions. Manipulator fits achieve R2 = 0.88 to 0.98 with slopes near unity, while vehicle surge, heave, and roll are reproduced with good fidelity under stronger coupling and noise. Median solver time is approximately 0.023 s per update, confirming online feasibility. A comparison against a fixed parameter model shows consistent reductions in MAE and RMSE across degrees of freedom. Results indicate physically plausible parameters and confidence intervals with near 100% coverage, enabling reliable feedforward control and simulation in underwater environments.