An EKF-Based Fast Tube MPC Scheme for Moving Target Tracking of a Redundant Underwater Vehicle-Manipulator System

Abstract

This paper presents a control scheme for addressing the moving target grabbing problem by a kinematically constrained redundant underwater vehicle-manipulator system (UVMS) in the presence of unmodeled uncertainties, sensory detecting noises, and time-varying external disturbances. Our proposed control scheme is designed by a robust fast tube model predictive controller (FTMPC) with an extended Kalman filter (EKF) target observer. The proposed robust FTMPC consists of an online fast nominal MPC and an online ancillary nonlinear controller to overcome the uncertainties and disturbances of the UVMS. The problem of the UVMS kinematic redundancy and mechanical/physical constraints is translated into a part of online optimization of FTMPC. The feasibility and stability of the FTMPC are proved. The proposed EKF observer is used to estimate the moving target trajectory and overcome the sensory measurement (system) noises, which can eliminate the assumption that most prior existing research works need to know the tracking reference trajectory precisely. Finally, the effectiveness of the proposed control scheme is verified through a series of UWSim simulations by the 6-degree-of-freedom (DoF) vehicle and 5-DoF manipulator GIRONA500 UVMS.