状态变换卡尔曼滤波的进一步解释及应用  

Further explanation and application of state transformation extended Kalman filter

作　　者：王茂松 吴文启[1] 何晓峰 潘献飞[1] WANG Maosong;WU Wenqi;HE Xiaofeng;PAN Xianfei(College of Intelligence Science and Technology,National University of Defense Technology,Changsha 410073,China)

机构地区：[1]国防科技大学智能科学学院

出　　处：《中国惯性技术学报》2019年第4期499-504,509,共7页Journal of Chinese Inertial Technology

基　　金：国家自然科学基金项目（61573371,61773394）

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3 min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。Aiming at the coordinate-frame consistency of velocity error vector,the rationality of the state transformation extended Kalman filter(ST-EKF)is further explained and applied to initial alignment and integrated navigation of strapdown inertial navigation system.In order to improve the estimation accuracies of gyroscope and accelerometer bias errors,a unified 15-state filter is used in the initial alignment and integrated navigation process.Single-position alignment experiment shows that the heading-angle estimation error of traditional extended Kalman filter(EKF)would gradually diverge with time due to the fact that the error state is not fully observable under the vibration environment,while the estimation result of ST-EKF is relatively stable.The 3 min two-position alignment experiments for both the ring laser gyro IMU and the fiber optic gyro IMU show that the heading estimation accuracy of the ST-EKF is more than 45%higher than that of the traditional EKF under the same propagation and updating frequencies between EKF and ST-EKF.Ring laser gyro IMU/GNSS and MEMS-IMU/GNSS integrated navigation vehicle experiments show that the ST-EKF integrated navigation system has higher positioning accuracy when GNSS signals are disconnected at multiple locations.

关 键 词：状态变换卡尔曼滤波 捷联惯导 初始对准 组合导航 

分 类 号：U666.1[交通运输工程—船舶及航道工程]