摘要
在小波变换对陀螺信号处理的基础上提出了用三层小波包消减阈值算法,处理陀螺仪的状态信号,小波包分解得到各个节点的分解系数。通过消减算法使得信号中大于阈值的瞬态高频部分置零,对消减阈值处理后的节点进行重构以消除信号中的噪声和周期分量。软阈值函数通过小波包分解的节点系数均值和方差来构建。仿真实验与已取得很好效果的强制降噪的重构信号和相应的功率谱进行比较,结果表明小波包软阈值降噪比强制降噪在随机漂移补偿上提高了13.1%,零偏改善了0.0526(°)/h。
Based on the processing of gyro＇s signal by wavelet transform, an algorithm using three-layer wavelet packet to de-noise threshold was proposed which was used to process the gyro＇s state signal. The decomposition coefficient of each node was got from the decomposition of third-layer wavelet packet. We let the high-fi＇equency transient part （greater than threshold） to be zero by a elimination algorithm, reconstructed the nodes to eliminate the noise and period components, the soft threshold function was constructed using the mean and variance of node coefficients decomposed by wavelet packet, the coefficients of the third nodes. Compared the forced de-noising method with the simulation using the proposed method, the results show that the.proposed method improves random drift compensation by 13. 1%, and reduced zero drift by 0.0526 （°）/h.
出处
《中国惯性技术学报》
EI
CSCD
2007年第5期602-605,共4页
Journal of Chinese Inertial Technology
关键词
小波变换
零偏
光纤陀螺仪
软阈值
信号去噪
wavelet transform
bias
fiber optical gyro
soft-threshold
de-noising
