航空兵器2026，Vol.33Issue(2)：74-80,7.DOI:10.12132/ISSN.1673-5048.2025.0143
基于Allan方差模型的GNSS/SINS组合导航自适应UKF算法
Adaptive UKF Algorithm for GNSS/SINS Integrated Navigation Based on Allan Variance Model
摘要
Abstract
The GNSS/SINS integrated navigation system is a typical nonlinear system,so the non-linear filtering methods is one of the effective ways to improve its filtering performance.UKF(Unscented Kalman Filter)algorithm is a common and important algorithm for nonlinear filtering of GNSS/SINS integrated navigation system,which requires that the system noise and measurement noise must be accurately known,otherwise the filtering accuracy will be reduced or even diverged.The appropriate adaptive filtering method to estimate the measurement noise variance in real time is the core of solving the aforementioned problem.An adaptive UKF algorithm(ALAUKF)for integrated navigation system is proposed based on Allan variance analysis in gyro noise modeling.Firstly,based on the algorithm flow of the UKF(Unscented Kalman Filter)for the integrated navigation system,and using the current measurement information and the state vector prediction values provided by the UKF,a sample sequence vector for the Allan variance analysis method was constructed.Then,based on the Allan variance analysis model,a measurement noise variance estimation model was proposed,and an forgetting factor model was constructed to improve the estimation accuracy of measurement noise variance.Subsequently,the ALAUKF algorithm for the GNSS/SINS integrated navigation system was proposed.Finally,the UKF,ALAUKF algorithms and the advanced variational Bayesian adap-tive UKF(VBAUKF)algorithm were applied to the nonlinear model of the GNSS/SINS integrated navi-gation system for experimental verification.The experimental results show that,when the measure-ment noise variance is unknown,compared with VBAUKF,ALAUKF can more accurately estimate various changes in the measurement noise variance(Fig.3 and Fig.4)and significantly improve the filtering accuracy of navigation parameters(Tab.1);and compared with UKF,ALAUKF can greatly enhance the filtering accuracy of the integrated navigation system(Tab.1 and Fig.2).Overall,com-pared with the VBAUKF algorithm,the ALAUKF algorithm has the advantage of a simpler algorithm,because it has fewer dynamic parameters.This has provided a solution approach for the research on adaptive filtering methods for nonlinear GNSS/SINS integrated navigation systems,and has enriched the category of adaptive filtering for nonlinear integrated navigation systems.关键词
组合导航/Allan方差分析法/测量噪声均方差估计/自适应滤波Key words
integrated navigation/Allan variance analysis method/measurement noise MSE esti-mation/adaptive filtering分类
信息技术与安全科学引用本文复制引用
刘思明,林雪原,乔玉新..基于Allan方差模型的GNSS/SINS组合导航自适应UKF算法[J].航空兵器,2026,33(2):74-80,7.基金项目
国家自然科学基金项目(62171402) （62171402）
山东省自然科学基金项目(2016ZRA06068) （2016ZRA06068）
