Vessel steering control using generalized ellipsoidal basis function based fuzzy neural networks

Abstract

This paper contributes to vessel steering control system design via the Generalized Ellipsoidal Function Based Fuzzy Neural Network (GEBF-FNN) method. Based on vessel motion dynamics and Nomoto model, a vessel steering model including dynamical K and T parameters dependent on initial forward speed and required heading angle is proposed to develop a novel dynamical PID steering controller including dynamical controller gains to obtain rapid and accurate performance. The promising GRBF-FNN algorithm is applied to dealing with the identification of dynamical controller gains. Typical steering maneuvers are considered to generate data samples for training the GEBF-FNN based dynamical steering controller while the prediction performance is checked by series of steering commands. In order to demonstrate the effectiveness of the proposed scheme, simulation studies are conducted on benchmark scenarios to validate effective performance. © 2012 Springer-Verlag.