Abstract
It is extremely difficult to estimate the parameters for the equation of ship roll motion, as compared to the other six degrees of freedom ship motion. Previously, Haddara et al. presented a method for estimating the non-linear term in the equation of motion by using a conventional neural network. In this study, numerical experiments were performed by estimating the non-linear term in the equation of ship roll motion by using the Self-Generating Radial Basis Function network developed by Katayama et al.
