摘要
为提高陀螺飞轮系统的标定与辨识精度进而保证姿态测量性能,对其信号去噪方法进行研究以便从复杂噪声中提取标定辨识所需的有用信息.首先基于噪声产生机理对陀螺飞轮信号的噪声特性进行分析.其次,在不影响信号低频特性的前提下,采用传统低通滤波(LPF)方法对信号进行预处理以抑制高频周期噪声.然后应用经验模态分解方法(EMD)对LPF预处理后的信号进行分解,根据信号和各模态分量的概率密度函数的相似性度量,给出一种模态判定准则.在此基础上,结合现有阈值滤波方法,提出一种EMD/LPF混合去噪方案.结果表明:所提出的基于相似性度量的模态判定准则在不同信噪比条件下均能够实现分界点的准确判定;将所研究的去噪方法分别应用于标准测试信号和陀螺飞轮实测信号;所提出的混合去噪方法相比于现有去噪方法更为有效.
In an effort to improve the calibration accuracy and ensure the rate sensing performance of gyrowheel,a suitable signal denoising method to extract the useful information from the complicated noise is essential.First,according to the causes of noise,the noise characteristics of gyrowheel were analyzed.Second,the traditional low-pass filter(LPF)was introduced to attenuate the high-frequency periodic noises without influencing the useful low-frequency characteristics.Then the empirical mode decomposition(EMD)method was applied to the filtered signal,and a criterion for selecting relevant intrinsic mode functions(IMF)was presented by using the similarity measurement between probability density functions of IMFs.Combining with the existing thresholding-based denoising technique,a hybrid EMD/LPF denoising strategy was proposed.Simulation results show that the proposed criterion for selecting relevant IMFs was always effective under different SNR conditions.Meanwhile,the proposed method was applied to standard test signals and real signals,and results show the effectiveness and superiority of the hybrid denoising method.
作者
赵昱宇
赵辉
霍鑫
姚郁
ZHAO Yuyu;ZHAO Hui;HUO Xin;YAO Yu(School of Astronautics,Harbin Institute of Technology,Harbin 150001,China)
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2020年第4期1-6,共6页
Journal of Harbin Institute of Technology
基金
国家自然科学基金(61427809,61773138)。
关键词
陀螺飞轮
信号去噪
经验模态分解
本征模态函数
巴氏距离
gyrowheel
signal denoising
empirical mode decomposition
intrinsic mode function
Bhattacharyya distance
作者简介
赵昱宇(1989-),女,博士研究生;赵辉(1971-),男,教授,博士生导师;姚郁(1963-),男,博士生导师,“长江学者”特聘教授;通信作者:姚郁,yaoyu@hit.edu.cn。
