Sigma点Kalman滤波在惯性导航初始对准中的性能评估  

Performance evaluation of sigma point Kalman filterfor SINS initial alignment

作　　者：张召友 郝燕玲[1] 张鑫[1] 

机构地区：[1]哈尔滨工程大学自动化学院,哈尔滨150001

出　　处：《中国惯性技术学报》2010年第6期639-644,共6页Journal of Chinese Inertial Technology

基　　金：国家自然科学基金项目(60775001)

摘　　要：捷联惯性导航系统的大失准角初始对准过程具有非线性特性,Sigma点卡尔曼滤波算法可用于解决大失准角的对准问题。为提高Sigma点卡尔曼滤波的数值稳定性并降低算法复杂度,推导了简化的方根形式的滤波算法。为评估不同Sigma点卡尔曼滤波在初始对准中的性能,引入了一种适用于姿态角全为大失准的欧拉角误差模型。最后对不同Sigma点卡尔曼滤波在初始对准中的应用进行了仿真,并对算法的精度与复杂度作了分析与总结。结果显示该算法可较好的处理大失准角情况下的对准问题,所推导的简化方根滤波算法在不影响对准精度的情况下可使算法的复杂度降低20%,并增强了数值稳定性,是一种较为实用的方法。In view that the initial alignment of strapdown inertial navigation system is nonlinear when with large misalignment angles,a sigma point Kalman filter（SPKF） is applied to deal with the nonlinear alignment problem,and a simplified square-root-based filter algorithm is derived for this sigma point Kalman filter to improve the numerical stability and reduce the computational complexity.To evaluate the performances of various sigma point Kalman filters in initial alignment,an Euler angle error model is introduced which can be used when three attitude angles are all large misalignment angles.The initial alignment simulations using various sigma point Kalman filters show that this filter can effectively deal with the nonlinear alignment problem.The simplified square-root-based filter algorithm decreases the computation burden by 20% without affecting the accuracy and enhances the stability of numerical calculation,showing that it is an efficient and practical method.

关 键 词：初始对准 Sigma点卡尔曼滤波 误差模型 复杂度 

分 类 号：U666.1[交通运输工程—船舶及航道工程]