Nonlinear model identification of a ship using recursive subspace methods

In this paper, we propose an online identification algorithm of ships for the dynamic positioning system. The ship model is first changed into Hammerstein form, then subspace identification method is used, with its simplicity and robustness, to perform the identification procedure. In order to make the identification algorithm more practical for online application, an effective updating strategy is proposed with Givens rotation matrices introduced. The proposed online identification method is tested on a DP-3 simulation system, with the assumption that the ship is operating under a calm sea condition, but the noise interference of the input/output measurement is considered. At last, simulation results showed the feasibility and effectiveness of the proposed identification method.