超球体平方根平淡卡尔曼滤波及其应用

Application of Hypersphere Square Root UKF on Initial Alignment of SINS

针对平淡卡尔曼滤波在捷联惯导大方位失准角下的初始对准中容易出现滤波数值不稳定以及计算量大的问题,提出了超球体平方根平淡卡尔曼滤波算法。该方法采用超球体分布采样点变换方法减少采样点数的求取,降低了算法的计算量;通过结合平方根滤波有效地克服了平淡卡尔曼滤波算法滤波数值不稳定的问题。仿真结果表明:该算法在保证初始对准滤波精度的前提下减小了计算量,提高了滤波效率。

While unscented Kalman filter（UKF）was used in initial alignment of large azimuth misalignment of strap-down inertial navigation system（SINS）,an hypersphere square root UKF algorithm was proposed to solve the problems of numerical instability and large calculation. This method reduced the amount of computation through spherical simplex unscented transformation（SSUT. The algorithm combined with square-root filter, the problem of numerical instability in UKF were solved efficiently. The simulation results showed that the improved algorithm could reduced the amount of calculation and had a better performance of filtering at the premise of en- suring the accuracy of alignment.