未知纬度条件下基于重力视运动与小波去噪的SINS自对准方法

基于惯性系的双矢量定姿方法选择惯性系中的两个重力视运动向量作为不共线矢量,解决了传统双矢量定姿方法在晃动基座条件下易受载体角运动干扰而无法实现对准的问题,但该方法仍需要精确的地理纬度信息以参与对准计算。针对未知纬度条件下的SINS抗晃动自对准问题,提出了一种基于重力视运动的三矢量自对准方法。该方法将初始对准问题归结为求解当前时刻导航系相对于初始时刻载体系的姿态矩阵问题,并利用矢量运算进行求解,仿真结果表明:加速度计随机测量噪声会映射为重力视运动随机噪声,降低对准精度;当加速度计随机噪声量级较大时,会带来对准计算失败。针对噪声问题,引入Daubechies(db4)小波进行5层分解来实现对重力视运动的降噪,并选择去噪后的重力视运动向量参与三矢量定姿解算,仿真结果表明:db4小波具有良好的去噪效果,基于小波去噪的三矢量自对准方法可以有效完成未知纬度条件下的SINS初始对准。

Self-alignment algorithm without latitude for SINS based on gravitational apparent motion and wavelet denoising

Double-vector attitude determination algorithm in inertial frame takes two gravitational apparent motion vectors as non-collinear vectors. Although this method solve the traditional algorithm’s problem that the information is susceptible to angular motion disturbance on swinging base, it still needs accurate latitude information to participate in alignment calculation. Aiming to fulfill the alignment for strapdown inertial navigation system without aided latitude information, a self-alignment method with three gravitational apparent motion vectors is designed. In this method, the alignment problem is attributed to solving the attitude matrix between current navigation frame and initial body frame and is solved with vector operation. Simulation results indicate that those random noises in the accelerator will be projected in gravitation apparent motion vectors and decrease the alignment accuracy, and even cause alignment failure when with large noise. For denoising, the daubechies (db4) wavelet is introduced to decompose gravitational apparent motions with 5 layers, and three denoised apparent motion vectors are selected to participate in the alignment. Simulation results indicate that the db4 owns excellent denoising effects and the alignment method with three apparent motion vectors and db4 in inertial frame can fulfill the alignment without aided latitude information.