传统参数辨识双位置对准方法绕方位轴转动90°构成第二位置，不是最优方案。改进方法绕方位轴转动180°构成最优双位置，并采用递推最小二乘进行参数辨识，计算量小，适合实时解算。采用光纤捷联惯导实测数据进行半物理仿真试验表明，改进方法10 min的方位对准精度优于1'，水平对准精度可提高一个数据量级，并且收敛速度得到明显改善。%The classic two-position parameter identification alignment method, which forms the second position by 90° rotation around the azimuth axis, is not optimal, while the improved method with 180° rotation is an optimal solution. The method identifies the concerned parameters by recursive least squares (RLS) algorithm, and its low computational complexity is very suitable for online solver. The effectiveness of proposed method is verified through semi-physical simulation experiments with the measured data of the fiber optic gyroscope (FOG) strapdown navigation system (SINS). The results show that after 10 min alignment, the azimuth misalignment error is less than 1′, and the level alignment accuracy is increased by one order of magnitude. Furthermore, the convergence speed can be improved obviously.
展开▼
