基于小波滤波的激光陀螺SINS晃动基座初始对准  

Wavelet used in initial alignment of SINS on a rocking base

作　　者：黄凤荣[1] 高峰[1] 付中泽[1] 蒋茂荣[1] 孙伟强[1] 

机构地区：[1]天津航海仪器研究所,天津300131

出　　处：《中国惯性技术学报》2014年第2期157-160,共4页Journal of Chinese Inertial Technology

基　　金：国防重点预研项目(51309030201)

摘　　要：激光陀螺捷联惯导系统在晃动基座上进行初始对准时,由于激光陀螺信号中存在着很大的随机噪声,导致对准时间变长,对准精度降低。为解决此问题,提出一种基于小波实时阈值滤波的预处理方法,选取了合适的小波基和分解层数,对陀螺信号进行小波预处理,然后利用滤波后的陀螺信号进行姿态粗对准,最后使用速度误差作为观测量进行卡尔曼滤波精对准。半实物仿真试验结果表明:本文提出的小波预处理方法在满足陀螺信号实时性的条件下,能够有效地减小激光陀螺信号中的各项随机噪声,利用预处理后的陀螺信号进行初始对准时,对准时间显著缩短,航向角在精对准7 min左右收敛,其1σ值在39"以内,在工程上有一定的参考价值。In initial alignment of RLG SINS on a stationary base, there are a lot of random noises in the RLG signals which may increase the alignment time and reduce the alignment precision. To solve this problem, a preprocessing method based on wavelet real-time threshold filtering is presented. The basic-wavelet and the proper level of wavelet decomposition are chosen. The gyro signal is preprocessed by using the wavelet. Then the coarse alignment is performed by using the filtered gyro signals. At last, velocity error matching and Kalman filtering are used to improve the precision of initial alignment. Experiment results show that the proposed wavelet preprocessing method can effectively reduce the RLG drift and satisfy the real-time demand. By using the preprocessed gyro signal, the initial alignment time is significantly reduced and the azimuth convergence can be accomplished in 7 min and the value of 1σ is within 39', which could provide some references in engineering application.

关 键 词：小波预处理 捷联惯导系统 初始对准 卡尔曼滤波 

分 类 号：U666.1[交通运输工程—船舶及航道工程]