摘要
针对动态条件下激光陀螺(RLG)随机噪声其非平稳和非正态分布的特性,提出了基于小波变换的RLG降噪方法,研究了小波去噪和Lipschitza指数奇异值检测的原理,给出了离散Db4小波滤波器的硬件实现方法,采用双正交静态小波基对RLG信号进行不同尺度的分解,通过不同小波基及低通滤波效果的比较,发现Db4小波具有较好的去噪性,其分解和重建滤波器系数也十分简单,验证了小波变换在陀螺仪信号去噪中的可行性,在用Db4滤波后的数据进行导航解算时,发现导航误差有很大的减小,研究成果对导航系统精度的提高有着重要的参考价值。
According to the properties of non-stationary and non-normal distribution of random drift in ring laser gyro （RLG）, it is difficult to eliminate the noise by traditional method. A method of denoising based on wavelet in RLG was employed. The theory for wavelet denoising and checking oddity of Lipschitza-exponent was researched. The hardware realization method of Db4 was designed. The residual signal from the gyroscope was decomposed by the different scale biorthogonal stationary wavelet base. By the comparison of filtering effect passing different wavelet base and low passing filter, the effect of Db4 was advantaged. The decomposing and reconstructed coefficient of this filter is relatively simple. The feasibility of wavelet denoising was validated in signal-process of RLG. The navigation error was greatly reduced when using data passing Db4 filter. The results of study are useful to the improvement of navigation system precision.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2008年第3期716-719,750,共5页
Journal of System Simulation
关键词
激光陀螺
RLG
小波去噪
奇异值检测
导航误差
Ring Laser Gyroscope
RLG
wavelet denoisie
checking oddity
navigation error
