基于DSP的EMD方法实现技术研究
|
摘要
本文主要研究了基于DSP的EMD方法实现技术。
传统非平稳信号处理方法存在不足和缺陷。短时傅立叶变换依赖于传统的傅立叶谱分析，所以限制于假定待分析数据是分段平稳的这个条件。Wigner-Ville分布克服了短时傅立叶分析的部分缺点，具有较高的分辨率，但存在严重的交叉干扰项。目前处理非平稳信号最好方法之一的小波分析也存在一些缺点。经验模态分解方法(EMD方法)是近年来出现的一种分析非线性、非平稳信号的新方法。使用这种方法可以将复杂的数据信号分解为有限数量的本征模函数(IMF)，对IMF分量进行Hilbert变换可以获得信号的时频分布。Huang提出的这个方法是通过Hilbert-Huang变换(HHT)工具包实现的。
鉴于目前国内外对EMD方法的研究多限于软件方面，课题尝试结合DSP和PCI技术来实现EMD方法，其优点和意义如下：
1)采用硬件系统来实现EMD方法可以方便地实现现场在线数据采集和分析，这在一些要求实时处理或准实时处理的场合尤为重要；
2)利用DSP强大的信号处理能力来实现EMD方法，可以充分发挥其高速和高效的优点：
3)可以充分利用PCI总线高传输速度、大地址空间等优点。
针对EMD方法的实现过程及DSP和PCI的特点，本文探讨了研制采集处理卡实现EMD方法的技术问题。论文所完成的工作如下：
1．实现了板卡的PCI总线接口，利用PCI总线将数据直接传到系统内存，有效解决数据的实时传输和存储，为信号的实时处理提供方便。
2．实现了输入信号的高速同步采集。所采集数据存放在DSP的片内RAM。
3．运用DSP实现采集数据的EMD分解。利用一种简易的镜像延拓处理信号的边缘效应，将存入片内RAM中的数据进行EMD分解，将每次分解所得的IMF分量传至PC机。
4．PC机端应用程序实现IMF分量的存储和显示工作。
本文作为一个有益的尝试，为EMD方法硬件实现的进一步探索和研究积累了一些经验。
The implementation technology of HMD method based on DSP is researched in this paper.
Traditional non-stationary data processing methods have some disadvantages. Short-time Fourier transform relies on the traditional Fourier spectral analysis, so it is limited to the assumption of the data to be piecewise stationary. The shortcoming of Wigner-Ville is the severe cross terms as indicated by the existence of negative power for some frequency ranges. So far one of the best available methods to process non-stationary data is wavelet analysis. However it has many problems. Empirical Mode Decomposition method (EMD) is a new method for analysing nonlinear and non-stationary data, with which any complicated data set can be decomposed into a finite and often small number of Intrinsic Mode Functions (IMF) that admit well-behaved Hilbert transforms. Thus the energy-frequency-time distribution of the signal is constructed. This method developed by Huang et al is implemented through HHT (Hilbert-Huang Transform) toolbox.
Because domestic and oversea research on EMD method at present is mainly focused on software, implementation of EMD method based on DSP and PCI is discussed in the paper. It has following advantages:
1) It's convenient to sample and analyze on-line signal through implementing EMD method by hardware, especially in some situations requiring realtime processing and quasi-realtime processing;
2) It's easy to show DSP's high-speed and high-efficiency as using its powerful signal processing ability to implement EMD method;
3) Rapid data transmission and other high performance of PCI local bus can be fully used.
According to the realizing procedure of EMD method and features of DSP and PCI, technologies of developing a signal acquisition/processing card and implementing EMD method is discussed. The paper has fulfill following jobs:
1. Implementation of PCI interface of this card. Data is transmitted to system memory directly with PCI local bus, which solves the realtime transmission and storage of data and also provides convenience for signal realtime processing.
2. Implementation of data acquisition with high speed and synchronization. The data is stored in on-chip RAM of DSP.
3. EMD method is implemented by DSP. A simple mirror-extension method is taken to sovle the end extending issue in EMD method. And data in on-chip RAM is processed by the approach, and each IMF is transmitted to PC.
4. The application on PC serves as the implementation of storage and display of IMF results.
The paper provides some experiences for future research on EMD implementation by hardware.
传统非平稳信号处理方法存在不足和缺陷。短时傅立叶变换依赖于传统的傅立叶谱分析，所以限制于假定待分析数据是分段平稳的这个条件。Wigner-Ville分布克服了短时傅立叶分析的部分缺点，具有较高的分辨率，但存在严重的交叉干扰项。目前处理非平稳信号最好方法之一的小波分析也存在一些缺点。经验模态分解方法(EMD方法)是近年来出现的一种分析非线性、非平稳信号的新方法。使用这种方法可以将复杂的数据信号分解为有限数量的本征模函数(IMF)，对IMF分量进行Hilbert变换可以获得信号的时频分布。Huang提出的这个方法是通过Hilbert-Huang变换(HHT)工具包实现的。
鉴于目前国内外对EMD方法的研究多限于软件方面，课题尝试结合DSP和PCI技术来实现EMD方法，其优点和意义如下：
1)采用硬件系统来实现EMD方法可以方便地实现现场在线数据采集和分析，这在一些要求实时处理或准实时处理的场合尤为重要；
2)利用DSP强大的信号处理能力来实现EMD方法，可以充分发挥其高速和高效的优点：
3)可以充分利用PCI总线高传输速度、大地址空间等优点。
针对EMD方法的实现过程及DSP和PCI的特点，本文探讨了研制采集处理卡实现EMD方法的技术问题。论文所完成的工作如下：
1．实现了板卡的PCI总线接口，利用PCI总线将数据直接传到系统内存，有效解决数据的实时传输和存储，为信号的实时处理提供方便。
2．实现了输入信号的高速同步采集。所采集数据存放在DSP的片内RAM。
3．运用DSP实现采集数据的EMD分解。利用一种简易的镜像延拓处理信号的边缘效应，将存入片内RAM中的数据进行EMD分解，将每次分解所得的IMF分量传至PC机。
4．PC机端应用程序实现IMF分量的存储和显示工作。
本文作为一个有益的尝试，为EMD方法硬件实现的进一步探索和研究积累了一些经验。
The implementation technology of HMD method based on DSP is researched in this paper.
Traditional non-stationary data processing methods have some disadvantages. Short-time Fourier transform relies on the traditional Fourier spectral analysis, so it is limited to the assumption of the data to be piecewise stationary. The shortcoming of Wigner-Ville is the severe cross terms as indicated by the existence of negative power for some frequency ranges. So far one of the best available methods to process non-stationary data is wavelet analysis. However it has many problems. Empirical Mode Decomposition method (EMD) is a new method for analysing nonlinear and non-stationary data, with which any complicated data set can be decomposed into a finite and often small number of Intrinsic Mode Functions (IMF) that admit well-behaved Hilbert transforms. Thus the energy-frequency-time distribution of the signal is constructed. This method developed by Huang et al is implemented through HHT (Hilbert-Huang Transform) toolbox.
Because domestic and oversea research on EMD method at present is mainly focused on software, implementation of EMD method based on DSP and PCI is discussed in the paper. It has following advantages:
1) It's convenient to sample and analyze on-line signal through implementing EMD method by hardware, especially in some situations requiring realtime processing and quasi-realtime processing;
2) It's easy to show DSP's high-speed and high-efficiency as using its powerful signal processing ability to implement EMD method;
3) Rapid data transmission and other high performance of PCI local bus can be fully used.
According to the realizing procedure of EMD method and features of DSP and PCI, technologies of developing a signal acquisition/processing card and implementing EMD method is discussed. The paper has fulfill following jobs:
1. Implementation of PCI interface of this card. Data is transmitted to system memory directly with PCI local bus, which solves the realtime transmission and storage of data and also provides convenience for signal realtime processing.
2. Implementation of data acquisition with high speed and synchronization. The data is stored in on-chip RAM of DSP.
3. EMD method is implemented by DSP. A simple mirror-extension method is taken to sovle the end extending issue in EMD method. And data in on-chip RAM is processed by the approach, and each IMF is transmitted to PC.
4. The application on PC serves as the implementation of storage and display of IMF results.
The paper provides some experiences for future research on EMD implementation by hardware.
引文
[1] Huang N E, Shen Z, Long S R, et al. The Empirical Mode Decomposition and the Hilbert Spectrum for Nonlinear and Non-stationary Time Series Analysis. Proc R Soc Lond A, 1998,454:899-995
[2] Huang N E, Shen Z, Long S R. A new view of nonlinear water waves: the Hilbert spectrum. Ann Rev Fluid Mech, 1999,31:417-457
[3] Huang, N. E., Long, S. R. & Shen, Z. The mechanism for frequency downshift in nonlinear wave evolution. Adv. Appl. Mech. 1996,32: 59-111．
[4] Huang N E, Chern CC, Huang K, et al. A new spectral representation of earthquake data: Hilbert spectral analysis of station TCU129, Chi-Chi, Taiwan, 21 September 1999． Bulletin Of The Seismological Society Of America, Oct2001, 91 (5) : 1310-1338
[5] Huang W, Shen Z, Huang NE, Fung YC. Engineering analysis of biological variables: An example of blood pressure over 1 day. Proceedings Of The National Academy of Sciences of The United States of America. APR 1998,95 (9) : 4816-4821
[6] Echeverria J C, Crowe J A, Woolfson M S, Hayes-Gill B R. Application of empirical mode decomposition to heart rate variability analysis. Medical & Biological Engineering & Computing. 2001, 39: 471-479
[7] Liang H, Lin Z, McCallum RW. Artifact reduction in electrogastrogram based on empirical mode decomposition method. Medical & Biological Engineering & Computing. JAN 2000,38 (1) : 35-41
[8] Komm RW, Hill F, Howe R. Empirical mode decomposition and Hilbert analysis applied to rotation residuals of the solar convection zone. Astrophysical Journal. SEP 2001,558 (1) : 428-441
[9] Vasudevan K, Cook FA. Empirical mode skeletonization of deep crustal seismic data: Theory and applications. Journal Of Geophysical Research-Solid Earth. APR 10 2000,105(84) : 7845-7856
[10] Loh CH, Wu TC, Huang NE. Application of the empirical mode decomposition-Hilbert spectrum method to identify near-fault ground-motion characteristics and structural responses. Bulletin Of The Seismological Society Of America. OCT 2001,91 (5) : 1339-1357
[11] Xun Zhu, Zheng Shen, Stephen D.E, et al. Gravity wave characteristics in the middle atmosphere derived from the empirical mode decomposition method. Journal of Geophysical Research, 1997,Vol. 102, No. D14: 16545-16561
[12] Montesinos ME, Munoz-Cobo JL, Perez C. Hilbert-Huang analysis of BWR neutron detector signals: application to DR calculation and to corrupted signal analysis. Annals Of Nuclear Energy. APR 2003, 30 (6) : 715-727
[13] Zhao JinPing, Huang DaJi. Mirror extending and circular spline function for empirical mode decomposition method. Journal of Zhejiang University (Science). 2001,2(No.3) : 247-252
[14] Yu Bo, Ma XJ. A New Method for the Analysis of Non-Stationary Nonlinear Vibration Signal and its Use in Machine Fault Diagnostics [R], Proceedings of International Conference on Vibration Engineering. 1998: 668-671．
[15] Benny Cheng. Data Analysis with the Empirical Mode Decomposition Method and the Hilbert Spectra. http://lochness.jpl.nasa.gov/EMD/emd.html.
[16] TMS320VC5402 fixed-point digital signal processor. Texas Instruments. 2000
[17] TMS320C54x Assembly Language Tools User's Guide. Texas Instruments.2001
[18] TMS320C54x DSP Reference Set, Volume 1: CPU and Peripherals. Texas Instruments. 2001
[19] TMS320C54x DSP Reference Set, Volume 2: Mnemonic Instruction Set. Texas Instruments. 2001
[20] TMS320C54x DSP Reference Set, Volume 4: Applications Guide. Texas Instruments. 1996
[21] TMS320C54x DSP Reference Set, Volume 5: Enhanced Peripherals. Texas Instruments. 1999
[22] TMS320C54x Optimizing C/C++ Compiler User's Guide. Texas Instruments. 2001
[23] TMS320C54x C Source Debugger User's Guide. Texas Instruments. 1994
[24] TMS320C54x Code Generation Tools Getting Started Guide. Texas Instruments. 1995
[25] Code Composer Studio Getting Started Guide. Texas Instruments. 2001
[26] TMS320C54x Code Composer Studio Tutorial. Texas Instruments. 2001
[27] PC1SIG, PCI Local Bus Specification, Revision 2, 1999
[28] AMCC PCI Products Data Book. 1998
[29] KRF-Tech WinDriver V4． Developer's Guide.2000
[30] G. Rilling, P. Flandrin, P. Goncalves. On Empirical Mode Decomposition and its Algorithms. IEEE-EURASIP Workshop on Nonlinear Signal and Image Processing NSIP-03, Grado (I). June 2003
[31] L．科恩[美]．时—频分析：理论与应用(白居宪译)．西安：西安交通大学出版社，1998
[32] 张贤达，保铮．非平稳信号分析与处理．北京，国防工业出版社，1998
[33] 姚天任，孙洪．现代数字信号处理．武汉：华中理工大学出版社，1999
[34] 张雄伟．DSP芯片的原理与开发应用．北京：电子工业出版社，2000
[35] 王念旭．DSP基础与应用系统设计．北京：北京航空航天大学出版社，2001
[36] 戴明桢，周建江．TMS320C54X DSP结构原理及应用．北京：北京航空航天大学出版社，2001
[37] 刘益成．TMS320C54x DSP应用程序设计与开发．北京：北京航空航天大学出版社，2002
[38] 盖强．局域波时域分析方法的理论研究与应用：[博士学位论文]．大连：大连理工大学，2001
[39] 邓拥军．EMD方法的改进：[硕士学位论文]．青岛：青岛海洋大学，2000
[40] 邓拥军，王伟，钱成春等．EMD方法及Hilbert变换中边界问题的处理．科学通报，2001．46(3)：257～263
[41] 赵进平．异常事件对EMD方法的影响及其解决方法研究．青岛海洋大学学报，2001．11：805～814
[42] 张海勇，马孝江，蔬强．一种新的时频分析方法．火力与指挥控制，2000．25(3)：39～42
[43] 赵光龙，基于TI—C54X的信号采集／处理PCI卡研制：[硕士学位论文]．成都：电子科技大学，2001
[44] 周文建．基于PCI总线的数据采集卡PnP驱动程序的研究：[硕士学位论文]．成都：电子科技大学硕士学位论文，2001
[45] 张宇斌．基于PCI总线的数字信号处理系统的软件开发：[硕士学位论文]．南京：南京航空航天大学，2002
[46] 王小兵．基于PCI总线的数字信号处理系统的硬件实现：[硕士学位论文]．南京：南京航空航天大学，2002
[47] 王长清．基于计算机PCI总线的雷达数据采集系统的研制：[硕士学位论文]．南京：南京理工大学，2002
[48] 王彤．基于DSP的带有USB接口的数据采集及处理电路的设计：[硕士学位论文]．杭州：浙江大学，2002
[49] 郑国轴．基于DSP技术的数据采集系统软件技术研究：[硕士学位论文]．杭州：浙江大学，2002
[50] 李鉴．PCI系统总线及其应用．现代电子技术，2002．1(135)：76～79
[51] 吴赟，张蕴玉，胡修林．PCI总线接口设计及其专用芯片应用．电子工程师，2002．28(2)：50～52
[52] 解璞，苏群星，谷宏强．PCI总线及其接口设计．计算机工程，2002．3：282～284
[53] 张军，修明磊，王莹．PCI总线接口芯片S5920及其应用．国外电子元器件，2002．4：64～67
[54] 张景涛，张弓．TMS320VC54系列EPROM的加载实现．国外电子元器件，2002．3：34～36
[55] 庄健敏，余松煜．TMS320C5x的混合语言编程，电子技术．2000．2：41～43
[56] 杨俊峰，何浩，武杰．多通道同步高速数据采集系统及其Windows程序设计．电子技术，2002．9：56～59
[2] Huang N E, Shen Z, Long S R. A new view of nonlinear water waves: the Hilbert spectrum. Ann Rev Fluid Mech, 1999,31:417-457
[3] Huang, N. E., Long, S. R. & Shen, Z. The mechanism for frequency downshift in nonlinear wave evolution. Adv. Appl. Mech. 1996,32: 59-111．
[4] Huang N E, Chern CC, Huang K, et al. A new spectral representation of earthquake data: Hilbert spectral analysis of station TCU129, Chi-Chi, Taiwan, 21 September 1999． Bulletin Of The Seismological Society Of America, Oct2001, 91 (5) : 1310-1338
[5] Huang W, Shen Z, Huang NE, Fung YC. Engineering analysis of biological variables: An example of blood pressure over 1 day. Proceedings Of The National Academy of Sciences of The United States of America. APR 1998,95 (9) : 4816-4821
[6] Echeverria J C, Crowe J A, Woolfson M S, Hayes-Gill B R. Application of empirical mode decomposition to heart rate variability analysis. Medical & Biological Engineering & Computing. 2001, 39: 471-479
[7] Liang H, Lin Z, McCallum RW. Artifact reduction in electrogastrogram based on empirical mode decomposition method. Medical & Biological Engineering & Computing. JAN 2000,38 (1) : 35-41
[8] Komm RW, Hill F, Howe R. Empirical mode decomposition and Hilbert analysis applied to rotation residuals of the solar convection zone. Astrophysical Journal. SEP 2001,558 (1) : 428-441
[9] Vasudevan K, Cook FA. Empirical mode skeletonization of deep crustal seismic data: Theory and applications. Journal Of Geophysical Research-Solid Earth. APR 10 2000,105(84) : 7845-7856
[10] Loh CH, Wu TC, Huang NE. Application of the empirical mode decomposition-Hilbert spectrum method to identify near-fault ground-motion characteristics and structural responses. Bulletin Of The Seismological Society Of America. OCT 2001,91 (5) : 1339-1357
[11] Xun Zhu, Zheng Shen, Stephen D.E, et al. Gravity wave characteristics in the middle atmosphere derived from the empirical mode decomposition method. Journal of Geophysical Research, 1997,Vol. 102, No. D14: 16545-16561
[12] Montesinos ME, Munoz-Cobo JL, Perez C. Hilbert-Huang analysis of BWR neutron detector signals: application to DR calculation and to corrupted signal analysis. Annals Of Nuclear Energy. APR 2003, 30 (6) : 715-727
[13] Zhao JinPing, Huang DaJi. Mirror extending and circular spline function for empirical mode decomposition method. Journal of Zhejiang University (Science). 2001,2(No.3) : 247-252
[14] Yu Bo, Ma XJ. A New Method for the Analysis of Non-Stationary Nonlinear Vibration Signal and its Use in Machine Fault Diagnostics [R], Proceedings of International Conference on Vibration Engineering. 1998: 668-671．
[15] Benny Cheng. Data Analysis with the Empirical Mode Decomposition Method and the Hilbert Spectra. http://lochness.jpl.nasa.gov/EMD/emd.html.
[16] TMS320VC5402 fixed-point digital signal processor. Texas Instruments. 2000
[17] TMS320C54x Assembly Language Tools User's Guide. Texas Instruments.2001
[18] TMS320C54x DSP Reference Set, Volume 1: CPU and Peripherals. Texas Instruments. 2001
[19] TMS320C54x DSP Reference Set, Volume 2: Mnemonic Instruction Set. Texas Instruments. 2001
[20] TMS320C54x DSP Reference Set, Volume 4: Applications Guide. Texas Instruments. 1996
[21] TMS320C54x DSP Reference Set, Volume 5: Enhanced Peripherals. Texas Instruments. 1999
[22] TMS320C54x Optimizing C/C++ Compiler User's Guide. Texas Instruments. 2001
[23] TMS320C54x C Source Debugger User's Guide. Texas Instruments. 1994
[24] TMS320C54x Code Generation Tools Getting Started Guide. Texas Instruments. 1995
[25] Code Composer Studio Getting Started Guide. Texas Instruments. 2001
[26] TMS320C54x Code Composer Studio Tutorial. Texas Instruments. 2001
[27] PC1SIG, PCI Local Bus Specification, Revision 2, 1999
[28] AMCC PCI Products Data Book. 1998
[29] KRF-Tech WinDriver V4． Developer's Guide.2000
[30] G. Rilling, P. Flandrin, P. Goncalves. On Empirical Mode Decomposition and its Algorithms. IEEE-EURASIP Workshop on Nonlinear Signal and Image Processing NSIP-03, Grado (I). June 2003
[31] L．科恩[美]．时—频分析：理论与应用(白居宪译)．西安：西安交通大学出版社，1998
[32] 张贤达，保铮．非平稳信号分析与处理．北京，国防工业出版社，1998
[33] 姚天任，孙洪．现代数字信号处理．武汉：华中理工大学出版社，1999
[34] 张雄伟．DSP芯片的原理与开发应用．北京：电子工业出版社，2000
[35] 王念旭．DSP基础与应用系统设计．北京：北京航空航天大学出版社，2001
[36] 戴明桢，周建江．TMS320C54X DSP结构原理及应用．北京：北京航空航天大学出版社，2001
[37] 刘益成．TMS320C54x DSP应用程序设计与开发．北京：北京航空航天大学出版社，2002
[38] 盖强．局域波时域分析方法的理论研究与应用：[博士学位论文]．大连：大连理工大学，2001
[39] 邓拥军．EMD方法的改进：[硕士学位论文]．青岛：青岛海洋大学，2000
[40] 邓拥军，王伟，钱成春等．EMD方法及Hilbert变换中边界问题的处理．科学通报，2001．46(3)：257～263
[41] 赵进平．异常事件对EMD方法的影响及其解决方法研究．青岛海洋大学学报，2001．11：805～814
[42] 张海勇，马孝江，蔬强．一种新的时频分析方法．火力与指挥控制，2000．25(3)：39～42
[43] 赵光龙，基于TI—C54X的信号采集／处理PCI卡研制：[硕士学位论文]．成都：电子科技大学，2001
[44] 周文建．基于PCI总线的数据采集卡PnP驱动程序的研究：[硕士学位论文]．成都：电子科技大学硕士学位论文，2001
[45] 张宇斌．基于PCI总线的数字信号处理系统的软件开发：[硕士学位论文]．南京：南京航空航天大学，2002
[46] 王小兵．基于PCI总线的数字信号处理系统的硬件实现：[硕士学位论文]．南京：南京航空航天大学，2002
[47] 王长清．基于计算机PCI总线的雷达数据采集系统的研制：[硕士学位论文]．南京：南京理工大学，2002
[48] 王彤．基于DSP的带有USB接口的数据采集及处理电路的设计：[硕士学位论文]．杭州：浙江大学，2002
[49] 郑国轴．基于DSP技术的数据采集系统软件技术研究：[硕士学位论文]．杭州：浙江大学，2002
[50] 李鉴．PCI系统总线及其应用．现代电子技术，2002．1(135)：76～79
[51] 吴赟，张蕴玉，胡修林．PCI总线接口设计及其专用芯片应用．电子工程师，2002．28(2)：50～52
[52] 解璞，苏群星，谷宏强．PCI总线及其接口设计．计算机工程，2002．3：282～284
[53] 张军，修明磊，王莹．PCI总线接口芯片S5920及其应用．国外电子元器件，2002．4：64～67
[54] 张景涛，张弓．TMS320VC54系列EPROM的加载实现．国外电子元器件，2002．3：34～36
[55] 庄健敏，余松煜．TMS320C5x的混合语言编程，电子技术．2000．2：41～43
[56] 杨俊峰，何浩，武杰．多通道同步高速数据采集系统及其Windows程序设计．电子技术，2002．9：56～59