摘 要
本次毕业设计的主要内容是应用经验模式分解（EMD）及其改进算法对脑电波信号进行分析，EMD是一种用于分析非线性和非平稳性时间序列的自适应数据处理和挖掘算法，该方法通过筛选过程将信号分解为若干本征模态函数IMF，每个IMF分量包含了原始信号在不同尺度上的局部特征。其次，结合机器学习算法，对经由EMD算法处理之后的信号进行分类识别。
本文首先引入了瞬时频率的概念，论述了IMF的基本概念和EMD分解算法原理，并且分析了EMD算法的特点，结合IMF给出了边际谱和Hilbert谱的物理含义，对整个 EMD 时频分析理论进行了详细的论述。对经验模式分解及其扩展算法进行技术调研，熟悉脑电波的常规处理方法，完成重要外文文献的翻译；期间按照节点撰写周志；研究脑电波信号的非线性分解方法，完成脑电波信号的特征提取流程设计；设计方案能综合考虑社会、健康、安全、法律、文化以及环境等因素。
关键词：经验模态分解；时频分析；HHT谱
ABSTRACT
The main content of this graduation project is to apply empirical pattern decomposition (EMD) and its improved algorithm to analyze EEG signals. EMD is an adaptive data processing and mining algorithm for analyzing nonlinear and non-stationary time series. The method decomposes the signal into a number of eigenmode functions IMF, each IMF component contains the local characteristics of the original signal on different scales through the screening process. Secondly, combined with the machine learning algorithm, the signals processed by the EMD algorithm are classified and recognized.
In this paper, the concept of instantaneous frequency is introduced, the basic concept of IMF and the principle of EMD decomposition algorithm are discussed, and the characteristics of EMD algorithm are analyzed. Combining with IMF, the physical meaning of marginal spectrum and Hilbert spectrum is given. The whole theory of EMD time-frequency analysis is discussed in detail. The empirical mode decomposition and its extended algorithms were investigated, familiar with the conventional processing methods of brain waves, and the translation of important foreign literature was completed, and the weekly records were written according to the nodes during the period. The nonlinear decomposition method of EEG signal is studied to complete the design of feature extraction flow of EEG signal, and the design scheme can consider the society comprehensively. Social, health, safety, legal, cultural and environmental factors.
Key words: empirical mode decomposition; time-frequency analysis; HHT spectrum
目 录
第1章 绪论 1
1.1 课题的研究背景和意义 1
1.2 EMD 方法的提出 3
1.3 EMD方法的研究发展概况 4
1.4 论文结构与安排 5
第2章 EMD时频分析的基本理论 7
2.1 EMD方法的基本概念 7
2.2 EMD方法的基本原理 8
2.3 基于EMD的HHT时频分析 12
2.4 EMD算法存在的问题及改进 13
2.5 本章小结 15
第3章 EMD 时频分析方法的应用 16
3.1 EMD 时频分析仿真及性能比较 16
3.2 EMD在脑电信号趋势提取中的应用 19
3.3 EMD 时频分析在奇异脑电信号检测中的应用 21
3.4 本章小结 24
第4章 基于EMD的脑电信号特征提取及其模式识别设计及仿真 25
4.1 传统的脑电信号特征和EMD脑电特征 25
4.2 EMD 的二进脑电特征特性及仿真 26
4.3 基于EMD的脑电特征方法及仿真 28
4.4本章总结 30
结 论 31
参考文献 32
致 谢 34
附 录 35
