Employing NMPC scheme utilizing improved cubature Kalman filter for an AUV system

Abstract

This paper deals with the diving-plane controller design problem for a class of torpedo-type Autonomous Underwater Vehicles (AUVs). Complex hydrodynamic characteristics of AUVs are truly onerous to model and control. AUVs exhibit peculiar nonlinear behaviour and have a high degree of coupling. As a result, they suffer from many practical limitations, such as a lack of hovering capability, limited payload-carrying capacity, poor pitch and roll motion stability, etc. To design an effective and reliable diving-plane control scheme for AUVs, this paper develops a new Stochastic Nonlinear Model Predictive Control (SNMPC) scheme relying on an improved Cubature Kalman Filtering (ICKF) algorithm. This scheme enables an AUV to reach the desired depth underwater and keep hovering at that depth. The ICKF algorithm estimates both the states and plant parameters to update the control law, which helps to reducethe effect of measurement noise and input disturbances. An in-depth simulation case study has been conducted in the presence of external disturbances, noises and plant-parameter variations to demonstrate the usefulness of the proposed ICKF-based NMPC scheme. The results are further analysed based on 'root mean square error' to show the controller's performance.