改进型卡尔曼算法解算小型弹箭姿态角研究

Research for Using Improved Kalman Algorithm to Solve Small Missiles Attitude Angle

弹箭飞行姿态参数的准确获取是实现精确高效打击敌方的重要前提。其中最重要的关键环节是姿态角的算法解算,决定着最后的精确导航。卡尔曼滤波算法是组合导航中最常用的解算算法,然而仍不能满足精度要求。为了进一步提高解算精度,提出通过BP神经网络对误差建模,进而修正卡尔曼滤波增益矩阵系数的改进算法。实验证明,应用于高速旋转弹姿态角测试,改进后的卡尔曼算法较传统卡尔曼算法精度提高了3°左右。

Acquiring the parameters of the flying projectiles’attitude is the important precise of the efficient attack against the enemy. The most important key segment is the algorithm of the posture corner,which decides to the last precise navigation. Kalman filter algorithm is common among the traditional posture corner measuring algorithm,but the precision of which is not still meet the requirements. Aim to improving the accuracy,the paper proposes a improved algorithm based on the BP nerve net which builds a error model,then the gain matrix coefficient of the Kalman filter is modified. The experiment result shows the accuracy of the improved algorithm is better than the traditional way of raising two degrees,while the improved algorithm applies to measuring the stance of the speed rotating projectiles.