Adaptive Controller to Minimize Ship Steering Propulsion Losses: Time Series Modeling and Statistical Analysis Results

Abstract

Abstract Statistical time-series and multivariable regression analysis techniques are employed to investigate the properties of a multivariable model for ship/steering and seaway disturbance dynamics. Confidence intervals for the model parameter values to be identified by least-squares in an adaptive steering autopilot scheme are calculated. Residual analysis techniques are used to verify the validity of models via data from ship/steering/seaway simulations. The sufficiency of a firstorder multivariable time-series model within a self-tuning adaptive control algorithm framework is shown to be justified by the results obtained.