基于SVD的改进抗差UKF算法及在组合导航中的应用
发布时间：2018-07-13 22:00
【摘要】：针对GPS/INS组合导航中因观测异常导致系统状态先验信息矩阵失去对称正定性,及传统等价权函数抗差算法易遇到病态矩阵,引起滤波性能下降的问题,提出一种基于奇异值分解的改进抗差UKF算法.该算法克服了先验协方差矩阵负定性变化,通过判断矩阵病态性实现智能选取抗差策略.最后利用车载实测数据进行验证,所得结果表明,SVD-UKF导航解精度稍优于EKF算法,改进的抗差策略能够极大减弱单独、连续以及混合的观测异常对导航解的影响,提高了导航解精度和可靠性.
[Abstract]:In view of the loss of symmetric positive definiteness of system state priori information matrix due to abnormal observation in GPS / ins integrated navigation system, the traditional equivalent weight function robust algorithm is prone to encounter ill-conditioned matrix, which leads to the degradation of filtering performance. An improved robust UKF algorithm based on singular value decomposition (SVD) is proposed. The algorithm overcomes the negative qualitative change of the priori covariance matrix and intelligently selects the robust strategy through the ill-condition of the judgment matrix. Finally, the experimental data are used to verify that the accuracy of SVD-UKF navigation solution is slightly better than that of EKF algorithm, and the improved robust strategy can greatly reduce the influence of single, continuous and mixed observation anomalies on the navigation solution. The accuracy and reliability of navigation solution are improved.
【作者单位】： 中国矿业大学国土环境与灾害监测国家测绘局重点实验室;中国矿业大学环境与测绘学院;
【基金】：新世纪优秀人才支持计划项目(NCET-13-1019) 江苏省高校优势学科建设工程项目(SZBF 2011-6-B35) 江苏省普通高校研究生科研创新计划项目(CXLX13 944)
【分类号】：TN96.2
[Abstract]:In view of the loss of symmetric positive definiteness of system state priori information matrix due to abnormal observation in GPS / ins integrated navigation system, the traditional equivalent weight function robust algorithm is prone to encounter ill-conditioned matrix, which leads to the degradation of filtering performance. An improved robust UKF algorithm based on singular value decomposition (SVD) is proposed. The algorithm overcomes the negative qualitative change of the priori covariance matrix and intelligently selects the robust strategy through the ill-condition of the judgment matrix. Finally, the experimental data are used to verify that the accuracy of SVD-UKF navigation solution is slightly better than that of EKF algorithm, and the improved robust strategy can greatly reduce the influence of single, continuous and mixed observation anomalies on the navigation solution. The accuracy and reliability of navigation solution are improved.
【作者单位】： 中国矿业大学国土环境与灾害监测国家测绘局重点实验室;中国矿业大学环境与测绘学院;
【基金】：新世纪优秀人才支持计划项目(NCET-13-1019) 江苏省高校优势学科建设工程项目(SZBF 2011-6-B35) 江苏省普通高校研究生科研创新计划项目(CXLX13 944)
【分类号】：TN96.2
【相似文献】
相关期刊论文 前4条
1 赵大志;唐海云;钟苏川;罗懋康;;基于UKF的反辐射导弹对移动辐射源的跟踪[J];信息与电子工程;2009年05期
2 徐t，
本文编号：2120876
本文链接：https://www.wllwen.com/kejilunwen/wltx/2120876.html
