Model-Based Dynamic Positioning of Underwater Robotic Vehicles: Theory and Experiment

This paper addresses the trajectory tracking problem for the low-speed maneuvering of fully actuated underwater vehicles. It is organized as follows. First, a brief review of previously reported control studies and plant models is presented. Second, an experimentally validated plant model for The Johns Hopkins University Remotely Operated Underwater Vehicle (JHUROV) is reviewed. Third, the stability of linear proportional-derivative (PD) control and a family of fixed and adaptive model-based controllers is examined analytically and demonstrated with numerical simulations. Finally, we report results from experimental trials comparing the performance of these controllers over a wide range of operating conditions. The experimental results corroborate the analytical predictions that the model-based controllers outperform PD control over a wide range of operating conditions. The exactly linearizing model-based controller is outperformed by its nonexactly linearizing counterpart. The adaptive controllers are shown to provide reasonable online plant parameter estimates, as well as velocity and position tracking consistent with theoretical predictions-providing good velocity tracking and, with the appropriate parameter update law, position tracking. The effects of reference trajectory, "bad" model parameters, feedback gains, adaptation gains, and thruster saturation are experimentally evaluated. To the best of our knowledge, this is the first reported comparative experimental study of this class of model-based controllers for underwater vehicles.