Abstract
The identification of dynamic models for ships is essential in marine and maritime engineering, as it facilitates accurate prediction and control of vessel behavior. Traditional ship system identification methods involve identifying a few parameters from the dynamic model. Only simplified linearized models have utilized metaheuristic techniques for ship identification recently. This paper describes a new way to use the Artificial Protozoa Optimization (APO) algorithm to make an Autoregression with Exogenous Input (ARX) model for the full ship dynamics work better. The adaptive foraging, dormancy, and reproduction behaviors of protozoa gave rise to the APO algorithm, which is a strong metaheuristic framework for parameter optimization. By leveraging the APO, we aim to enhance the accuracy of ship dynamics identification and its explainability. Evaluation of the model on the provided test data demonstrates that the APO-optimized ARX model provides excellent identification compared to results from existing methods, achieving a mean R2score of 0.9437, thus providing a reliable and efficient tool for ship maneuvering identification.