Buchumschlag

A Suppression Method for Random Errors of IFOG Based on the Decoupling of Colored Noise-Spectrum Information

Gespeichert in:
Veröffentlicht in:Micromachines vol. 16, no. 8 (2025), p. 963-984
1. Verfasser: Liang Zhe
Weitere Verfasser: Zhang, Zhili, Zhou Zhaofa, Li, Hongcai, Zhao Junyang, Tian Longjie, Duan, Hui
Veröffentlicht:
MDPI AG
Schlagworte:
Navigation systems
Random errors
Fiber optic gyroscopes
Strapdown inertial navigation
Model forms
Wavelet transforms
Fog
Equations of state
Modelling
Decoupling
Random walk
Linear systems
Noise
Autoregressive moving average
Methods
Algorithms
Kalman filters
Time series
Optics
Markov analysis
Inertial navigation
Online-Zugang:Citation/Abstract
Full Text + Graphics
Full Text - PDF
Tags: Tag hinzufügen
Keine Tags, Fügen Sie das erste Tag hinzu!
Abstract:In high-precision inertial navigation systems, suppressing the random errors of a fiber-optic gyroscope is of great importance. However, the traditional rule-based autoregressive moving average modeling method, when applied in Kalman filtering considering colored noise, presents inherent disadvantages in principle, including inaccurate state equations and difficulties in state dimension expansion. To this end, the noise characteristics in the fiber-optic gyroscope signal are first deeply analyzed, a random error model form is clarified, and a new model-order determination criterion is proposed to achieve the high-precision modeling of random errors. Then, based on the effective suppression of the angle random walk error of the fiber-optic gyroscope, and combined with the linear system equation of its colored noise, an adaptive Kalman filter based on noise-spectrum information decoupling is designed. This breaks through the principled limitations of traditional methods in suppressing colored noise and provides a scheme for modeling and suppressing fiber-optic gyroscope random errors under static conditions. Experimental results show that, compared with existing methods, the initial alignment accuracy of the proposed method based on 5 min data of fiber-strapdown inertial navigation is improved by an average of 48%.
ISSN:2072-666X
DOI:10.3390/mi16080963
Quelle:Engineering Database