基于矩阵卡尔曼滤波的捷联惯导初始对准算法

Initial alignment algorithm for SINS based on state matrix Kalman filter

针对捷联惯导系统大失准角晃动基座条件下的初始对准问题,提出了一种基于矩阵卡尔曼滤波的抗干扰自对准算法。该方法将传统大失准角非线性对准问题,简化为确定初始时刻姿态的线性矩阵卡尔曼滤波估计问题。借鉴惯性系REQUEST算法,将重力矢量在惯性系下的投影作为量测,利用K矩阵在对准过程中为常值特性,以其作为待估计的状态可避免系统模型误差和初始误差的影响,同时避免了传统方法对失准角大、小的假设,也不再区分粗、精对准过程,适用于任意姿态、无初值条件下的对准。在发动机振动及外界扰动条件下进行了四个方位的对准试验,试验表明,对于导航级惯导系统,算法可在5 min内完成初始对准且统计方位均方差小于3'(1σ),略优于传统算法。

An anti-disturbance initial self-alignment algorithm using vector observations and the matrix Kalman filter (MKF) are presented for strapdown inertial navigation system (SINS). Employing the gravity vector projection in inertial frame as measurements provides an optimal estimation in a state matrix Kalman filter framework, and the nonlinear alignment problem for large misalignment angles is equally transformed into a linear optimization of K-matrix. The initial K-matrix related to the initial attitude is a constant matrix, and the kinematics equation is linear and precise. The above characteristics yield a linear MKF that eliminates the usual linearization procedure and is less sensitive to initial estimation errors. The proposed algorithm naturally fits the random misalignment angles and does not need the coarse alignment stage. A four orientation repeated alignment experiment under the engine vibration and external random disturbances is carried out, which demonstrates the efficiency of the MKF. The results show that the new algorithm can converge within 5 min for navigation-grade strapdown inertial navigation system (SINS), and the root mean square (1σ) of yaw error is less than 3′.