Application of an offline identification algorithm for adjusting parameters of a modular manoeuvring mathematical model
Serge Sutulo, C. Guedes Soares
Abstract
The paper describes the continuation of studies carried out earlier in the application of system identification to the estimation of parameters of ship manoeuvring mathematical models. A new version of a versatile identification algorithm based on various metrics and global optimization schemes is described. The algorithm is tested on artificial training data obtained with an entirely different ship mathematical model. The presented results confirm sufficient effectiveness and robustness of the algorithm which is now considered as sufficiently matured for application to real-world experimental data. At the same time, the study demonstrates the importance of using sufficiently complete and detailed mathematical structures and of establishing appropriate and sufficiently large intervals for the model parameters’ variations.