Dynamics Modelling and Parameters Estimation for Nonlinear-Model Predictive Control of Underwater Robots

Abstract

Underwater robots are to be extensively used in various submerged missions. A few research works have been conducted to investigate the nonlinear dynamical behavior of such systems in ocean flows. In this paper, dynamics modeling and parameters estimation for these complicated systems are developed for using model-based predictive control. It is begun by presenting the nonlinear equations governing the dynamics of an underwater robot in an ocean open flow. Additionally, the movement of the diving and depth motion of the robot in such cases is analyzed. A method to estimate hydrodynamic and other parameters of an underwater robot is presented and compared with experimental data. This method consists of close-to-reality simulation using a CFD subsystem of the ANSYS software. Finally, these parameters are used to design a model-based controller. Numerical simulations are carried out to evaluate the performance of the controllers. Among all the model-based controllers, nonlinear Model Predictive Control (nonlinear-MPC) with constraints will be able to follow the setpoints within less time and less control effort. The experimental data confirm the simulation results, proving the underwater robot's capability for path-following in an ocean open flow.