捷联惯导( SINS)合成孔径雷达(SAR)组合导航系统中,SINS与SAR天线附加的惯性测量元件(IMU)之间的抗干扰、动态快速传递对准是一个研究难题.为了既不增加滤波器阶数,减小计算量,又能提高传递对准的速度和精度.在吸收现有滤波算法优点的基础上,提出了一种新的奇异值分解H∞联邦滤波算法.首先,建立了“速度+姿态+位置”匹配更新的传递对准模型,利用基于奇异值分解的H∞子滤波器进行滤波,得到系统状态的局部最优估计值；然后,通过联邦滤波器对局部最优估计值进行融合得到全局最优估计值.仿真结果表明,提出的滤波算法得到的失准角估计值能在200S内收敛,且分别稳定在3′、-5′和20′,性能明显优于H∞滤波和联邦滤波.新的滤波算法不但速度快,精度高,而且计算量小,抗干扰性好,为提高传递对准的精度提供了一种新方法和新途径.%In strapdown inertial navigation system(SINS)/synthetic aperture radar(SAR) integrated navigation systems, the anti-interference and dynamic rapid transfer alignment between SINS and the additional inertial measurement unit (IMU) of the antenna for S AR is a difficult problem. In order improve the speed and precision of transfer alignment without increasing the order of filter, and reduce the amount of calculation, the paper presents a new singular value decomposition H∞ federal filtering algorithm by absorbing the advantages of existed filtering methods. First, the model of transfer alignment for "velocity + attitude + position" matching update is established, and the robust filtering is carried on by the slave filter for H∞ filtering which is based on singular value decomposition to obtain the local optimum estimation of system states. Then, the global optimum estimation is derived based on the local optimum estimation fused by the federal filter. The simulation results demonstrate that the misalignment angles' estimations, which are obtained by the proposed filtering algorithm, are converged on 3', -5' and 20' in 200 s respectively, and the filtering performance is significantly superior to that of H∞ filtering and federal filtering. The new filtering algorithm has the characteristics of fast speed, high accuracy, small calculation amount and good anti-interference performance, providing an effective approach to improve the accuracy of transfer alignment.
展开▼
