表面肌電信號(hào)的壓縮感知稀疏表示分類研究

2.13萬(wàn)字
自己原創(chuàng)的畢業(yè)論文，僅在本站獨(dú)家出售，重復(fù)率低，推薦下載使用

摘要： 表面肌電信號(hào)(Surface Electromyographic, sEMG)的模式分類是智能多功能假肢控制的基本問(wèn)題。當(dāng)人體肌肉收縮時(shí)，通過(guò)安放在皮膚表面的電極可采集到肌肉相應(yīng)動(dòng)作電位，即表面肌電信號(hào)。表面肌電信號(hào)被廣泛地應(yīng)用于肌肉損傷診斷、康復(fù)醫(yī)學(xué)及體育運(yùn)動(dòng)等方面的研究，其中一項(xiàng)重要應(yīng)用為利用表面肌電信號(hào)控制假肢。本論文基于壓縮感知和稀疏表示理論，研究前臂肌肉在展拳、握拳、內(nèi)旋、外旋四種運(yùn)動(dòng)模式下產(chǎn)生的表面肌電信號(hào)的特征提取和多類別模式識(shí)別方法。首先介紹了表面肌電信號(hào)的基礎(chǔ)知識(shí)和壓縮感知的理論框架及特點(diǎn)。然后基于壓縮感知理論對(duì)多類別表面肌電信號(hào)進(jìn)行稀疏表示，將肌電信號(hào)測(cè)試樣本表示為訓(xùn)練樣本集的過(guò)完備字典稀疏線性組合，使用隨機(jī)測(cè)量矩陣獲取測(cè)試樣本降維特征量和稀疏表示感知矩陣，應(yīng)用正交匹配追蹤法求取肌電信號(hào)測(cè)試樣本的稀疏解，由冗余誤差最小值確定目標(biāo)歸屬類，實(shí)現(xiàn)對(duì)展拳、握拳、內(nèi)旋、外旋四種運(yùn)動(dòng)模式表面肌電信號(hào)的稀疏多分類識(shí)別。仿真研究表明基于壓縮感知理論的隨機(jī)矩陣降維映射特征提取不依賴于肌電信號(hào)模式類別，構(gòu)造簡(jiǎn)單，運(yùn)算快速，具有普適性； 稀疏表示分類法無(wú)需組合多個(gè)二分類器來(lái)實(shí)現(xiàn)多類肌電信號(hào)識(shí)別，且識(shí)別率較高。壓縮感知與稀疏表示理論在表面肌電信號(hào)模式分類中的應(yīng)用，能有效簡(jiǎn)化肌電信號(hào)特征提取與模式識(shí)別過(guò)程，為表面肌電信號(hào)模式識(shí)別研究提供了新的思路和方法。
關(guān)鍵字：表面肌電信號(hào)（sEMG) 模式識(shí)別 壓縮感知 正交匹配追蹤 稀疏性


Study on Classification of Surface EMG Signal Based on Compressive Sensing
Abstract： Surface electromyography (Surface Electromyographic, sEMG) pattern classification is the basic problem of multi function control artificial intelligence.When the human body muscle contraction, the skin surface can be placed on the electrode into the muscle action potential acquisition, namely the surface EMG signal.Surface EMG signal is widely used in the diagnosis of muscle injury, rehabilitation and sports and other aspects, one important application for the use of surface EMG signal controlled prosthesis.In this paper, based on the theory of compressed sensing and sparse representation, feature extraction and multi class pattern recognition method of surface electromyography of the forearm muscles in the exhibition fist, fist, pronation external rotation, four motion mode.The paper first introduces the basic knowledge of surface EMG signal and framework and characteristics of the compressed sensing theory.Then the compressed sensing theory are sparse on multiple categories of surface EMG signal based representation, the EMG signal test sample is expressed as a linear combination of the training sample set complete sparse, using the random measurement matrix for obtaining test sample reduction features and sparse representation of sensing matrix, and using orthogonal matching pursuit method for testing the sample dilute of the EMG signal, the minimum of the redundancy error target attributive in order to classify the four movement modes of surface EMG signal. Simulation results show that the random matrix theory of compressed sensing based on dimensionality reduction mapping feature extraction is not dependent on the classification of the EMG pattern , has the advantages of simple structure, fast calculation, universality.Sparse representation classification without the need to combine multiple classifiers to achieve two many kinds of EMG signal recognition, and the recognition rate is high.Compressed sensing and sparse representation theory applied in pattern classification of surface EMG, can effectively simplify the EMG signal feature extraction and pattern recognition process, and provides a new idea and method for the research of pattern recognition of surface EMG signal.

Key word：Surface Electromyographic(sEMG) Pattern Recognition Compressed Sensing Orthogonal Matching Pursuit Sparsity


目錄
摘要 I
Abstract II
第1章 緒論 - 1 -
1.1 研究背景和意義 - 1 -
1.2 國(guó)內(nèi)外的研究現(xiàn)狀 - 2 -
1.2.1 肌電信號(hào)的特征提取 - 2 -
1.2.2 肌電信號(hào)的模式分類 - 3 -
1.3 主要研究?jī)?nèi)容 - 3 -
1.4 小結(jié) - 4 -
第2章 肌電信號(hào)的特征提取和分類方法 - 5 -
2.1 表面肌電信號(hào)產(chǎn)生原理 - 5 -
2.2 肌電信號(hào)的特點(diǎn)簡(jiǎn)介 - 6 -
2.3 肌電信號(hào)的處理方法 - 7 -
2.3.1 肌電信號(hào)的常見(jiàn)分析方法 - 7 -
2.3.2 肌電信號(hào)的分類方法 - 8 -
2.4 小結(jié) - 9 -
第3章 壓縮感知基本理論 - 10 -
3.1 壓縮感知的背景 - 10 -
3.2 壓縮感知理論 - 11 -
3.2.1 壓縮感知理論簡(jiǎn)介 - 11 -
3.2.2 壓縮感知數(shù)學(xué)模型 - 13 -
3. 2.3 正交匹配追蹤法 - 13 -
3.3 壓縮感知理論的應(yīng)用和優(yōu)點(diǎn) - 14 -
3.3.1 壓縮感知的應(yīng)用 - 14 -
3.3.2 壓縮感知的優(yōu)點(diǎn) - 15 -
3.4 基于壓縮感知原理的肌電信號(hào)的特征提取和模式識(shí)別 - 16 -
3.4.1 肌電信號(hào)的稀疏表示特征提取 - 16 -
3.4.2 正交匹配追蹤法對(duì)信號(hào)進(jìn)行模式識(shí)別 - 17 -
3.5 小結(jié) - 17 -
第4章 表面肌電信號(hào)的壓縮感知模式分類 - 18 -
4.1 表面肌電信號(hào)的采集 - 18 -
4.2肌電信號(hào)的稀疏表示特征提取 - 18 -
4.2.1 肌電信號(hào)的稀疏表示 - 18 -
4.2.2 肌電信號(hào)的稀疏求解 - 19 -
4.2.3 隨機(jī)降維映射的稀疏表示分類法 - 19 -
4.3 MATLAB仿真 - 20 -
4.3.1 MATLAB仿真流程 - 2..