Nonlinear autoregressive model for space tracking ship's swaying data errors

Abstract

To get a thorough understanding of the properties and regulations of space tracking ship's swaying data errors so as to effectively eliminate their influences on measurement data, a nonlinear auto-regressive moving average model with exogenous inputs (NARMAX) has been developed by applying modern theories and the latest technologies. The process model of NARMAX was determined by orthogonal regressive method and the algorithm to determine noise model was also designed. After evaluating its fitting errors by comparing them with those of linear ARMA models, also by performing statistics and test for the redundant residuals, we can draw such conclusions as: The proposed NARMAX model can better fit ship swaying data errors, whose covariance and mean values prove to be much smaller than those processed by linear ARMA models, also it is more stationary and accurate, all of which validate that NARMAX can give a better description of the complicated characteristics and objective regulations of ship swaying data errors.