ESO-Based Robust Model Predictive Control for Undersea Vehicle Manipulator System

Abstract

In view of the complexity and uncertainty of the marine environment and the strong nonlinearity and coupling of the undersea vehicle manipulator system(UVMS), this paper proposed a robust model predictive control(RMPC) method based on extended state observer(ESO). First, a dynamics model was established based on the dynamics characteristics of UVMS, and a nominal model system was defined by ignoring uncertainties and disturbances. Then, a UVMS algorithm was designed for the nominal system. The uncertainties, disturbances, modeling errors, and other influencing factors of the original system were summarized into extended states, and an ESO was designed to estimate these factors. Furthermore, the factors were compensated based on the RMPC of the nominal model, so as to obtain the RMPC method applied to the UVMS system. Finally, it is demonstrated through simulation experiments that the ESO-based RMPC has good trajectory tracking performance and anti-disturbance capability.

Keywords

• undersea vehicle manipulator system
• robust model predictive control
• extended state observer
• trajectory tracking.