Gaussian Process-Augmented Model Predictive Contouring Control for 6-DoF Trajectory Tracking with a Holonomic AUV

TL;DR A 6-DoF model predictive contouring controller for a small holonomic underwater vehicle, with Gaussian-process residuals that correct the rotational dynamics. In a test tank it cuts the attitude error of the vehicle's field-proven PID controller by 47% on a lawnmower survey, and it follows an inspection spiral that the PID cannot.
What they do
Small holonomic AUVs such as MBARI’s MOLA can control position and attitude independently, which suits close-range inspection, but their hydrodynamics are hard to identify. The authors run a nonlinear model predictive contouring controller (MPCC) in all six degrees of freedom on MOLA and correct its nominal model with residual dynamics learned by Gaussian process (GP) regression. They describe it as the first real-time deployment of such a scheme on a holonomic AUV.
How it works
- Contouring control. The controller follows a spline path parameterized by arc length, treats progress along it as a decision variable, and penalizes lag error, contouring error and the attitude error against a SLERP-interpolated attitude reference.
- GP residuals. The nominal model uses BlueROV2 coefficients, since MOLA has no identified model. One GP per body velocity learns the mismatch from open-loop excitation data, sub-sampled to 100 points per GP.
- Real-time robustness. Uncertainty is propagated outside the solver iterations (zero-order robust optimization) to tighten constraints, so the problem keeps its nominal structure and runs at 25 Hz with a 40-step horizon.
Results
- In simulation, learning only the angular residuals lowers lawnmower attitude RMSE from 0.81° to 0.55°. Learning residuals for all six velocities roughly doubles position and attitude errors, because the linear residuals are mostly noise.
- On a simulated inspection spiral, attitude RMSE falls from a mean of 31° (nominal) to 4.5°.
- In a saltwater tank, attitude RMSE is 4.5° (PID), 3.2° (nominal MPCC) and 2.4° (GP-MPCC). The GP-MPCC’s position RMSE rises to 0.040 m from 0.033 m, and it uses 28% more control energy than the nominal MPCC.
- Both MPCC variants follow the inspection spiral in the tank; the PID moves between waypoints one degree of freedom at a time.
Limitations
The GPs are trained offline and do not adapt during long deployments. State-estimation uncertainty is not propagated, and the in-water tests compare five lawnmower runs per controller in a single tank.
Figures



Abstract
Small, holonomic Autonomous Underwater Vehicles (AUVs) enable close-range inspection of underwater structures from arbitrary viewpoints, requiring accurate position and attitude tracking under state and input constraints. Model Predictive Contouring Control (MPCC) naturally handles these requirements, but relies on an accurate dynamics model, which is difficult to obtain underwater. In this paper, we present a nonlinear MPCC in all six Degrees of Freedom (DoF) for a holonomic AUV and alleviate the inaccuracies in the dynamics by augmenting the nominal model employed with residual dynamics learned via Gaussian Process (GP) regression. The learned residual is embedded in the optimal control problem through a zero-order robust optimization scheme, which keeps the controller real-time capable at 25 Hz. A simulation ablation shows that residual selection is essential when designing the controller, and simulation and in-water experiments with the Multimodality, Observing, Low-cost, Agile (MOLA) AUV validate the resulting controllers against a nominal MPCC and a field proven Proportional-Integral-Derivative (PID) controller.
Summaries are written automatically from each paper's text. Figures and videos come from the paper and its project page; all credit goes to the authors.
