中文摘要：

针对SINS海上对准中速度等量测值的先验统计特性未知导致Cubature滤波（Cubature KF）的失准角估计值出现严重振荡的问题,提出变分贝叶斯Cubature滤波.该滤波方法通过近似计算状态变量和量测噪声方差的联合条件后验概率密度,在估计状态变量的同时,实时调整变分贝叶斯参数,估计和修正时变的量测噪声方差,减弱量测噪声方差统计特性对状态变量估计值的影响.半实物仿真结果表明,该方法能够减小Cuba-ture KF海上对准失准角估计值的振荡性,提高海上对准的精度.

英文摘要：

Due to the unknown velocity measurements statics,the misalignment angles of SINS（strapdown inertial navigation system）offshore alignment based on Cubature filter are fluctuated sharply.Thus,the variational Bayesian Cubature filter（Cubature KF）was proposed,which was deduced by approximation calculation of the conjugate posterior density probability of the state variables and the measurement noise variances.Therefore,through adjusting the variational Bayesian parameters,the time varying noise variances were estimated and modified,then the state estimation error was reduced.The hardware-in-the-loop simulation results show that compared with the stand Cubature KF,the proposed method provides better filter stability and accuracy for offshore alignment.