Date of Award

Fall 2023

Project Type


Program or Major

Mechanical Engineering

Degree Name

Doctor of Philosophy

First Advisor

May-Win Thein

Second Advisor

Bingxian Mu

Third Advisor

Se Young Yoon


Modeling and control of marine robotics have become a conspicuous research topic in the controlengineering community in the last decade. Remotely operated underwater vehicles (ROVs) are deployed and used in commercial, military, and scientific areas for many different purposes such as inspection and maintenance of submerged infrastructure, underwater surveillance/maritime, security, and ocean floor mapping most of which cannot be achieved effectively by human operators. However, to be able to operate these vehicles, an effective controller must be obtained, and an accurate dynamic mathematical model of the vehicle from which to design the controller is necessary. Although conventional modeling approaches are widely used to design a variety of motion controllers, these controllers are not effective against harsh and changing environmental conditions of open water missions, coastal zones, shallow water, or changing vehicle configurations, as these changes significantly affect model parameters. To provide an efficient adaptive controller, for example, real-time updates of model parameters are required. Here, nonlinear system identification techniques are considered effective approaches to rigorously model underwater vehicle dynamics. This research proposes a method to identify the dynamics of a small-scale uncrewed underwater vehicle with complex dynamics (e.g., tether dynamics) by using a system identification technique that incorporated Nonlinear Auto Regressive Moving Average with eXogenous Input (NARMAX) model structure via a "black-box" approach. Based on the experimentally obtained model, various motion controllers are designed. Experiments were conducted at the facilities of the University of New Hampshire’s Jere A. Chase Ocean Engineering Laboratory. It is shown that the developed nonlinear system identification method produced nonlinear models which can successfully represent vehicle dynamics. Validation of these models is provided via both numerical and experimental methods. Furthermore, it is shown that the models obtained by the proposed identification method can be used to design adaptive controllers and operational configuration changes of the vehicle can be compensated with the controllers designed. The effectiveness of the controllers is also shown via both numerical simulations and experiments.