摘要
针对舰载环境下,主/子惯导组合中主、子惯导间的挠曲变形问题,设计了H?滤波算法。算法将主、子惯导间的挠曲变形作为一种能量有限的量测噪声,通过H?滤波对不确定噪声的鲁棒性来保证滤波精度。在经典的"速度+姿态"匹配方法的基础上,仿真比较了:1)方案一,已知挠曲变形时的标准Kalman滤波算法;2)方案二,未知挠曲变形时的标准Kalman滤波算法;3)方案三,未知挠曲变形时的H?滤波算法。仿真误差曲线直观表明:基于H?滤波算法可以在未知挠曲变形时有效提高组合精度;统计结果表明:滤波方案三得到的姿态与速度的系统误差、随机误差均与理想条件下的方案一相当;相较于方案二,滤波方案三得到的姿态与速度的系统误差至少可减小40%,随机误差至少可减少15%。
A filter is designed to compensate for the flexure deformation between MINS（master inertial navigation system） and SINS（slave inertial navigation system） in the integration of M/SINS on ship board.In this algorithm,the deformation is considered to be a measurement noise with limited energy,and the filter accuracy is assured by the robust of filter to the uncertain noise.Based on the classical ＂velocity plus attitude＂ matching method,simulations are made to compare three schemes,they are： 1） standard Kalman filter with known accuracy deformation;2） standard Kalman filter with unknown deformation,and 3） filter with unknown deformation.Simulation result and the statistical data indicate that： 1） in the first and third schemes,the system and random errors of attitude and velocity are equivalent;2） but compared with the second scheme,the system and random errors by the third scheme are decreased at least about 40% and 15%.
出处
《中国惯性技术学报》
EI
CSCD
北大核心
2012年第1期74-78,共5页
Journal of Chinese Inertial Technology
基金
自然科学基金(61004125,60874092)
