APPLICATION OF THE EXTENDED KALMAN FILTERING TECHNIQUE TO SHIP MANEUVERING ANALYSIS

This thesis dealt with the application of a particular technique in systems identification, the Kalman statistical filter, to maneuvering analyses, determining the value of the hydrodynamic coefficients to the general equations of motion. A computer program was developed for use in this identification process. The system that the identification was applied to was the general class of surface vessels. The Mariner-class hull form was singled out for extensive analysis because of the availability of accepted values for the coefficients of these ships in the literature. The identification process was conducted over a variety of experimental conditions. The results indicate a capability for the program to identify the desired coefficients with reasonable accuracy - within five percent of the accepted true values for the individual coefficients. It was found that the best type of maneuver was one which generates a continuously varying input of the vessel's motion parameters, such as the sinusoidal maneuver. Additionally, the process was shown to be able to operate on noisy data containing a large amount of scatter. The new coefficient estimates can be refiltered on additional passes by the process over the same noisy data and thereby re-evaluated and updated to a new estimate. The results of this updating seems to depend upon the accuracy of the estimates obtained from the previous pass over the noisy data.

  • Corporate Authors:

    Massachusetts Institute of Technology

    Department of Ocean Engineering, 77 Massachusetts Avenue
    Cambridge, MA  United States  02139
  • Authors:
    • Lundblad, J G
  • Publication Date: 1975-1

Media Info

  • Pagination: 3 p.

Subject/Index Terms

  • TRT Terms: Maneuvering
  • Subject Areas: Marine Transportation; Vehicles and Equipment;

Filing Info

  • Accession Number: 00097504
  • Record Type: Publication
  • Source Agency: Massachusetts Institute of Technology
  • Report/Paper Numbers: MS Thesis
  • Files: TRIS
  • Created Date: Jul 24 1975 12:00AM