Ship Dynamics Identification Using Simulator and Sea Trial Data.

摘要

The Maritime Surface and Subsurface (MARS) Virtual Reality Simulator (VRS) has a role in determining the transfer of training. Determining the transfer of training requires knowledge of the ship dynamics. This paper identifies the ship dynamics for a CF Bay-Class ship as well as a simulated ship. Simulation and sea trial data are collected and used for a system's identification exercise. Sources of error came from the sea state, crew behaviours, and the Differential Global Positioning System. Despite the sources of error, the data were relatively clean and the identification exercise was able to proceed. The analysis produced a piecewise linear and continuous model description for the ship's dynamics, with numerical coefficients. The advantage of the model is that it is computationally simpler than the full six degree of freedom model, and still captures over 90% of the experimental variance. The algorithms developed in this paper may be implemented to generate a real time ship dynamics model that could be downloaded into a portable MARS VRS system for onboard training and rehearsal.