基于改进集合经验模态分解的陀螺信号降噪方法

吴倩; 刘昱

doi:doi:10.3788/LOP57.150601

激光与光电子学进展, 2020, 57 (15): 150601, 网络出版: 2020-08-04

基于改进集合经验模态分解的陀螺信号降噪方法下载： 966次

De-Noising Method for Gyroscope Signal Based on Improved Ensemble Empirical Mode Decomposition

吴倩刘昱 ^*

作者单位

天津大学微电子学院, 天津 300072

图 & 表

图 1. 信噪比为5 dB的Blocks信号经EEMD后的结果。(a)含噪信号;(b)IMF分量;(c)残余分量和部分IMF分量

Fig. 1. Results of Blocks signal obtaiend by EEDM with R_SN=5 dB. (a) Noisy signal; (b) IMF components; (c) residual component and partial IMF components

下载图片查看原文

图 2. EEMD-M原理框图

Fig. 2. Principle of proposed EEMD-M

下载图片查看原文

图 3. 信噪比为5 dB的Blocks信号经EEMD后各IMF分量与原信号的相关系数值

Fig. 3. Correlation coefficient of IMF components and original signal after EEMD for Blocks signal with R_SN=5 dB

下载图片查看原文

图 4. 原始信号及不同算法对陀螺信号的降噪结果(以z轴为例)。(a)原始噪声;(b)小波去噪;(c) EMD去噪;(d) DFA-EMD去噪;(e) EEMD去噪;(f)所提算法

Fig. 4. Original signal and de-noising results of gyroscope signal using different algorithms (taking z-axis as an example). (a) Original signal; (b) wavelet denoising; (c) EMD denoising; (d) DFA-EMD denoising; (e) EEMD denoising; (f) proposed algorithm

下载图片查看原文

图 5. 原始信号及不同算法降噪后信号的Allan方差结果

Fig. 5. Allan variance results of original signal and de-noising signals obtained by different algorithms

下载图片查看原文

表 1本文EEMD-M与其他几种算法对Blocks信号降噪的结果对比

Table1. Comparison of de-noising results for Blocks signal using proposed EEMD-M and some other algorithms

Signal	Input R_SN /dB	R_SN after denoising /dB
Signal	Input R_SN /dB	Wavelet	EMD	EMD-DFA	EEMD	Proposed method
	1	9.46	9.63	10.06	7.71	12.97
	2	10.32	9.81	10.79	8.74	13.50
	3	10.75	9.83	11.62	9.67	13.96
	4	11.50	9.14	12.41	10.69	14.48
	5	11.90	9.20	13.17	11.65	14.99
	6	12.72	13.87	13.78	12.59	15.41
	7	13.26	14.58	14.78	13.54	15.87
Blocks	8	14.06	15.25	14.69	14.43	16.50
	9	14.56	16.02	15.80	15.31	17.05
	10	15.14	16.62	14.25	16.18	17.61
	11	15.92	17.05	16.10	16.90	18.15
	12	16.71	17.56	14.40	17.79	18.92
	13	17.43	16.60	17.82	18.47	19.54
	14	18.30	17.03	15.50	19.12	20.06
	15	18.88	17.79	18.66	19.82	20.51

查看原文

表 2本文算法与其他几种算法对陀螺信号的降噪结果对比

Table2. Comparison of de-noising results of gyroscope signal using proposed algorithm and some other algorithms

Signal	Noisysignal	Waveletdenoising signal	EMD signal	DFA-EMDsignal	EEMDsignal	Signal obtained byproposed method
MSE	4.4075×10^-2	1.0777×10^-2	9.2160×10^-3	8.3660×10^-3	8.7990×10^-3	7.5510×10^-3
MAE	0.1672	0.0901	0.0868	0.0863	0.0866	0.0862
Maximum error	0.8537	0.3572	0.2764	0.2123	0.2348	0.1143

查看原文

吴倩, 刘昱. 基于改进集合经验模态分解的陀螺信号降噪方法[J]. 激光与光电子学进展, 2020, 57(15): 150601. Qian Wu, Yu Liu. De-Noising Method for Gyroscope Signal Based on Improved Ensemble Empirical Mode Decomposition[J]. Laser & Optoelectronics Progress, 2020, 57(15): 150601.

基于改进集合经验模态分解的陀螺信号降噪方法下载： 966次

图 1. 信噪比为5 dB的Blocks信号经EEMD后的结果。(a)含噪信号;(b)IMF分量;(c)残余分量和部分IMF分量

Fig. 1. Results of Blocks signal obtaiend by EEDM with R_SN=5 dB. (a) Noisy signal; (b) IMF components; (c) residual component and partial IMF components

图 2. EEMD-M原理框图

Fig. 2. Principle of proposed EEMD-M

图 3. 信噪比为5 dB的Blocks信号经EEMD后各IMF分量与原信号的相关系数值

Fig. 3. Correlation coefficient of IMF components and original signal after EEMD for Blocks signal with R_SN=5 dB

图 4. 原始信号及不同算法对陀螺信号的降噪结果(以z轴为例)。(a)原始噪声;(b)小波去噪;(c) EMD去噪;(d) DFA-EMD去噪;(e) EEMD去噪;(f)所提算法

Fig. 4. Original signal and de-noising results of gyroscope signal using different algorithms (taking z-axis as an example). (a) Original signal; (b) wavelet denoising; (c) EMD denoising; (d) DFA-EMD denoising; (e) EEMD denoising; (f) proposed algorithm

图 5. 原始信号及不同算法降噪后信号的Allan方差结果

Fig. 5. Allan variance results of original signal and de-noising signals obtained by different algorithms

表 1本文EEMD-M与其他几种算法对Blocks信号降噪的结果对比

Table1. Comparison of de-noising results for Blocks signal using proposed EEMD-M and some other algorithms

表 2本文算法与其他几种算法对陀螺信号的降噪结果对比

Table2. Comparison of de-noising results of gyroscope signal using proposed algorithm and some other algorithms

关于本站 Cookie 的使用提示

全站搜索

基于改进集合经验模态分解的陀螺信号降噪方法 下载： 966次

图 1. 信噪比为5 dB的Blocks信号经EEMD后的结果。(a)含噪信号;(b)IMF分量;(c)残余分量和部分IMF分量

Fig. 1. Results of Blocks signal obtaiend by EEDM with RSN=5 dB. (a) Noisy signal; (b) IMF components; (c) residual component and partial IMF components

图 2. EEMD-M原理框图

Fig. 2. Principle of proposed EEMD-M

图 3. 信噪比为5 dB的Blocks信号经EEMD后各IMF分量与原信号的相关系数值

Fig. 3. Correlation coefficient of IMF components and original signal after EEMD for Blocks signal with RSN=5 dB

图 4. 原始信号及不同算法对陀螺信号的降噪结果(以z轴为例)。(a)原始噪声;(b)小波去噪;(c) EMD去噪;(d) DFA-EMD去噪;(e) EEMD去噪;(f)所提算法

Fig. 4. Original signal and de-noising results of gyroscope signal using different algorithms (taking z-axis as an example). (a) Original signal; (b) wavelet denoising; (c) EMD denoising; (d) DFA-EMD denoising; (e) EEMD denoising; (f) proposed algorithm

图 5. 原始信号及不同算法降噪后信号的Allan方差结果

Fig. 5. Allan variance results of original signal and de-noising signals obtained by different algorithms

表 1本文EEMD-M与其他几种算法对Blocks信号降噪的结果对比

Table1. Comparison of de-noising results for Blocks signal using proposed EEMD-M and some other algorithms

表 2本文算法与其他几种算法对陀螺信号的降噪结果对比

Table2. Comparison of de-noising results of gyroscope signal using proposed algorithm and some other algorithms

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索

基于改进集合经验模态分解的陀螺信号降噪方法下载： 966次

Fig. 1. Results of Blocks signal obtaiend by EEDM with R_SN=5 dB. (a) Noisy signal; (b) IMF components; (c) residual component and partial IMF components

Fig. 3. Correlation coefficient of IMF components and original signal after EEMD for Blocks signal with R_SN=5 dB