SINS姿态估计的HoMM-UKF模型算法  

The Ho MM-UKF Algorithm of SINS'Attitude Estimation Model

作　　者：丁国强[1] 张铎[1] 熊明[1] 周卫东[2] 

机构地区：[1]郑州轻工业学院电气信息工程学院,河南郑州450002 [2]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001

出　　处：《郑州大学学报（工学版）》2017年第4期39-45,共7页Journal of Zhengzhou University（Engineering Science）

基　　金：国家自然科学基金资助项目(U1204603)

摘　　要：针对舰船捷联惯性导航系统姿态控制计算精度要求,提出一种高阶矩匹配UKF(high-order moment matching UKF,Ho MM-UKF)的SINS系统四元数模型姿态估计算法.在UKF迭代递推计算过程中利用高阶矩匹配方法计算系统状态参数的预测采样点集及其权值的概率分布平均偏态和峰值,使其精确逼近状态参数最优估计.采用四元数姿态建模方法构建新型SINS状态变量与噪声向量相关的姿态方程模型,利用伪观测向量构建观测噪声与四元数相关的观测方程模型,设计系统噪声方差分离计算算法开展系统噪声方差计算,引入拉格朗日乘子算法计算四元数估计均值,最后利用SINS四元数姿态估计模型对HoMM-UKF算法开展仿真试验研究.通过UKF算法、CDKF算法与HoMM-UKF算法对比,验证了HoMM-UKF算法计算精度高,并且算法计算量负担较小,计算效率较高.In order to improve the precision requirement about the attitude control of the strap-down inertial navigation system, the high order moment matching UKF （Higher-order Moment Matching UKF, HoMM- UKF） algorithm was proposed, that is to estimate the SINS＇ attitude parameters of based on its quaternion er- ror model. In the recursive calculation process, for accurately approximating computational purposes, it uses high order moment matching method to calculate the average skewness value and peak value of the predicted sampling points set and their weights of the system state parameters in the view of the probability distribution. Making use of attitude quaternion method, then onlinear quaternion error model was constructed, in which model the systemnoise vector depends on system state vector, meanwhile construct its measure equation whose measurement noise vector depends on quaternion measurement vector by pseudo observation vector method was constructed, the weighted average of estimated quaternion with Lagrangian operator was calculated, the system noise variance calculation with the system noise separation algorithm was carried out, and finallyconstruct the SINS＇ attitude estimation HoMM-UKF algorithms simulation on SINS attitude experiment platform was de- signed. It can be seen that HoMM-UKF algorithm＇ s calculation accuracy is higher than others and has better numerical stability, comparison of the UKF, and CDKF algorithms, and so the HoMM-UKF algorithm＇s feasi- bility and calculation accuracy is verified.

关 键 词：捷联惯性导航 姿态估计 高阶矩匹配 HoMM-UKF算法 Lagrangian乘子 

分 类 号：TP271[自动化与计算机技术—检测技术与自动化装置]