一种大安装偏差快速传递对准中的闭环卡尔曼滤波算法

In rapid transfer alignment with large physical misalignment angles, Kalman filter (KF) obtains biased estimation due to the model linearization error. To solve this problem, a transfer alignment approach based on closed-loop Kalman filter (CLKF) is presented. A coarse reference of the slave inertial navigation system (INS) is established by combining the estimated physical misalignment angles with the attitude of the master INS, which minimizes the misalignment angles of the slave INS and reduces the influence on the estimation accuracy of KF induced by the linearization error. Meanwhile, the slave INS state and the estimation of the misalignment angles are corrected with state feedback, which drives the slave INS reference body frame and the calculated body frame converging to the nominal body frame gradually. So that the whole system is always kept in a good linear state in the entire transfer alignment process. The simulation results show that the estimation accuracy of CLKF is equal to UKF and better than KF, which is not affected by the increase of system nonlinearities. Analogous to KF, the computational cost of CLKF is only 3% of UKF, which improves the estimation accuracy of the linear filter significantly and effectively avoids the problem of large computational burden caused by applying the nonlinear filter.