Performance Analysis of Marine Guidance Systematic Error Separation based on Linear Model
Xuanying Zhou, Zhengming Wang, Dong Li, Jiongqi Wang · 2018 IEEE 7th Data Driven Control and Learning Systems Conference (DDCLS) · 2018
As the guidance systematic errors of inertial missiles directly determine the guidance accuracy, error separation is a vital data processing problem. The key point of error separation is to find out a good parameter estimation method and to design a suitable estimation strategy according to the errors' physical characteristics. Based on the linear regression model of the guidance systematic error separation, this study gives the comparisons of four parameter estimation methods, which are Least Square Estimation(LSE), Bayesian estimation, Principal Component Analysis(PCA) and regularization method, and gives the simulations of PCA and Regularization method. Moreover, combining with the initial errors of sea-based missiles, we design two estimation strategies named the sorting strategy and the iteration strategy. The results illustrate that these two new strategies can separate more errors than the traditional overall strategy.