(1.9)--状态变换卡尔曼滤波的进一步解释及应用.pdfVIP

状态变换卡尔曼滤波的进一步解释及应用

摘要：针对速度误差定义的坐标系一致性，对状态变换卡尔曼滤波（ST-EKF）的合理性作出了进一步

的解释，并将其应用于捷联惯导的初始对准及组合导航。为提高陀螺仪、加速度计的零偏误差估计效

果，初始对准和组合导航过程采用统一的15状态滤波器。“单”位置精对准实验表明，由于在基座振

动的环境下误差状态不完全可观，传统卡尔曼滤波会有随时间增长航向角估计误差反而逐渐发散的情

况出现，而状态变换卡尔曼滤波的航向角估计结果相对稳定。激光陀螺IMU和光纤陀螺IMU的3min

双位置对准实验均表明，在同样的卡尔曼滤波预测和更新频率下，状态变换卡尔曼滤波的航向角估计

精度要比传统卡尔曼滤波提高45％以上。激光陀螺IMU/卫星和MEMS-IMU/卫星组合导航跑车实验均

表明，多处断开卫星信号的情况下，状态变换卡尔曼滤波组合导航系统具有更高的定位精度。

关键词：状态变换卡尔曼滤波；捷联惯导；初始对准；组合导航

中图分类号：U666.1文献标志码：A

Furtherexplanationandapplicationofstate

transformationextendedKalmanfilter

WANGMaosong,WUWenqi,HEXiaofeng,PANXianfei

(CollegeofIntelligenceScienceandTechnology,NationalUniversityofDefenseTechnology,Changsha410073,China)

Abstract:Aimingatthecoordinate-frameconsistencyofvelocityerrorvector,therationalityofthestate

transformationextendedKalmanfilter(ST-EKF)isfurtherexplainedandappliedtoinitialalignmentand

integratednavigationofstrapdowninertialnavigationsystem.Inordertoimprovetheestimationaccuracies

ofgyroscopeandaccelerometerbiaserrors,aunified15-statefilterisusedintheinitialalignmentand

integratednavigationprocess.Single-positionalignmentexperimentshowsthattheheading-angle

estimationerroroftraditionalextendedKalmanfilter(EKF)wouldgraduallydivergewithtimeduetothe

factthattheerrorstateisnotfullyobservableunderthevibrationenvironment,whiletheestimationresult

ofST-EKFisrelativelystable.The3mintwo-positionalignmentexperimentsforboththeringlasergyro

IMUandthefiberopticgyroIMUshowthattheheadingestimationaccuracyoftheST-EKFismorethan

45%higherthanthatofthetr