Thermally Induced Bias Drift Integrated Compensation for the IFOG Strapdown Inertial Navigation System

Abstract

The large thermally induced bias drift is the main factor affecting the performance of the interferometer fiber optic gyroscope (IFOG) in practical engineering applications. The thermally induced bias drift is investigated in order to improve the accuracy of IFOG in the strapdown inertial navigation system (SINS). A triaxial integrated linear multivariable model (TILMM) is presented based on experiments. The model is composed of the temperature gradient, polynomial temperature and temperature square. Due to the linear character of the model, model parameters are identified from the temperature and system outputs by least-square (LS) estimation methods. The drift data sets are collected to validate the effectiveness of this model. The rough experiment data is preprocessed by the proposed sliding window median absolute deviation (SWMAD) technology to identify outliers firstly, then the gradient inverse weighted (GIW) filter is used to initially lower the noise, and the singular value decomposition (SVD) is used to reduce the noise of the experiment data. The verification experiment of compensation is conducted and shows that the TILMM is effective, and SINS bias stability is improved in comparison with single-axis compensation.