Skip to Main Content
ASME Press Select Proceedings

International Conference on Information Technology and Computer Science, 3rd (ITCS 2011)

Editor
V. E. Muhin
V. E. Muhin
National Technical University of Ukraine
Search for other works by this author on:
W. B. Hu
W. B. Hu
Wuhan University
Search for other works by this author on:
ISBN:
9780791859742
No. of Pages:
656
Publisher:
ASME Press
Publication date:
2011

The linear auto regression (AR) model used for predict ships' movement, which with strong nonlinear characteristics, can not meet the precision need. To enhance the ship's seaworthiness capacity, a new prediction algorithm based on nonlinear auto regression model is presented to forecast roll motion accurately. The state-dependent autoregressive(SDAR) model and RBF(radial basis function) network are introduced firstly, then using RBF to map the functional coefficients of SDAR model, and the structure of the nonlinear parameter optimization method(SNPOM) is adopted to optimize parameters. The new model named RBF-NAR model, which has nonlinear approximation ability and auto regression characteristics, is applied to predict ships' rolling sequence. The comparison between AR and RBF-NAR of same order shows that higher accuracy can be achieved by the later one. Further more, RBFNAR costs less time than pure RBF network prediction with same precision requirement. The simulation results prove that the new model is more accurate and stabilizer than tradition models.

This content is only available via PDF.
Close Modal
This Feature Is Available To Subscribers Only

Sign In or Create an Account

Close Modal
Close Modal