In order to estimate the current state of ship motion and, based upon it, predict the short-term behavior of ship motion, the governing equation is obtained from the shape of typical ship-wave transfer function in the form of 2nd-order differential equation with constant coefficients. Then, so as to obtain high performance in the prediction of small-sized motion, a new algorithm is developed, in which prediction of ship motion in the stochastic environment is performed by the deterministic analysis. In the algorithm, wave input information is extracted from the estimated ship motion data and, based upon it, the ship motion prediction is performed by extrapolating the governing equation of ship motion.
Comparison of this algorithm with a typical Kalman filtering technique is made by simulations with a newly defined index.