基于MATLAB的膝关节振动信号（VAG）分析的简单模拟研究与实现

目标：

• 区分正常与病理性膝关节VAG信号：通过模拟信号生成、滤波、峰值检测与时频分析，判断信号是否异常。
• 关键指标：基于滤波后信号的峰值数量（>30则判定为异常）。

模块:

(1) 信号生成

• 正常信号：vag_normal = sin(2*pi*50*t) + 0.2*randn(size(t))

• 设计意图：50Hz正弦波模拟正常关节振动，叠加弱噪声（标准差0.2）。

• 病理性信号：vag_pathology = sin(2*pi*50*t) + 1.0*sin(2*pi*120*t) + 0.7*randn(size(t))
• 设计意图：在50Hz基础上添加120Hz高频成分（幅度1.0）和强噪声（标准差0.7），模拟病理特征。

(2) 滤波设计

• 滤波器类型：3阶巴特沃斯带通滤波器（20-250Hz）。

• 作用：保留VAG信号主要频段（20-250Hz），抑制低频肌电干扰和高频噪声。
• 实现：butter(3, [20 250]/(fs/2), bandpass) + filtfilt（零相位滤波）。
• 优点：filtfilt消除相位失真，确保峰值时间对齐。

(3) 峰值检测

• 正常信号参数：MinPeakHeight=0.5, MinPeakDistance=fs/20（50ms间隔）。
• 病理性信号参数：MinPeakHeight=0.3, MinPeakDistance=fs/50（20ms间隔）。

• 设计意图：病理性信号预期有更多密集高频峰值，降低高度阈值和最小间隔以捕获更多峰值。

(4) 异常判断

• 阈值：peak_threshold = 30（总峰值数 >30 判定为异常）。
• 逻辑验证：正常信号预期峰值较少，病理性信号因高频成分和噪声导致峰值增多。

% Define Sampling Frequency and Time Vectorfs = 1000; % Sampling frequency in Hzduration = 5; % Duration of signal in secondst = 0:1/fs:duration-1/fs; % Time vector

% Simulate Normal and Pathological VAG Signalsvag_normal = sin(2*pi*50*t) + 0.2*randn(size(t)); % Normal VAG: 50 Hz with noisevag_pathology = sin(2*pi*50*t) + 1.0*sin(2*pi*120*t) + 0.7*randn(size(t)); % Pathological: 50 Hz + 120 Hz with increased amplitude% Bandpass Filter Design (20-250 Hz)[b, a] = butter(3, [20 250]/(fs/2), bandpass); % 3rd order Butterworth bandpass filter

% Apply Bandpass Filtervag_normal_filtered = filtfilt(b, a, vag_normal); % Filter normal signalvag_pathology_filtered = filtfilt(b, a, vag_pathology); % Filter pathological signal% Detect Peaks in Filtered Signals% Adjust parameters to detect more peaks in pathological signal[peaks_normal, locs_normal] = findpeaks(vag_normal_filtered, MinPeakHeight, 0.5, MinPeakDistance, fs/20);[peaks_pathology, locs_pathology] = findpeaks(vag_pathology_filtered, MinPeakHeight, 0.3, MinPeakDistance, fs/50); % Lower height, closer peaks

% Define Abnormality Detection Thresholdpeak_threshold = 30; % Number of peaks threshold for abnormal detection% Abnormality Detection Logic (Based on Filtered Signals)is_abnormal_normal = length(peaks_normal) > peak_threshold; % True if abnormalis_abnormal_pathology = length(peaks_pathology) > peak_threshold; % True if abnormal% Generate Abnormality Status Stringsnormal_status = No Detection; % Default for normal signalif is_abnormal_normal    normal_status = Yes Detection;endpathological_status = No Detection; % Default for pathological signalif is_abnormal_pathology    pathological_status = Yes Detection;end% Visualizationfigure(Name, Knee Joint VAG Signal Analysis, NumberTitle, off);% Plot 1: Original Normal Signalsubplot(4, 2, 1);plot(t, vag_normal);xlabel(Time (s));ylabel(Amplitude);title(Original Normal VAG Signal);grid on;% Plot 2: Original Pathological Signalsubplot(4, 2, 2);plot(t, vag_pathology);xlabel(Time (s));ylabel(Amplitude);title(Original Pathological VAG Signal);grid on;% Plot 3: Filtered Normal Signal with Detected Peakssubplot(4, 2, 3);plot(t, vag_normal_filtered);hold on;plot(locs_normal/fs, peaks_normal, ro); % Mark peaksxlabel(Time (s));ylabel(Amplitude);title(Filtered Normal VAG Signal with Peaks);grid on;% Plot 4: Filtered Pathological Signal with Detected Peakssubplot(4, 2, 4);plot(t, vag_pathology_filtered);hold on;plot(locs_pathology/fs, peaks_pathology, ro); % Mark peaksxlabel(Time (s));ylabel(Amplitude);title(Filtered Pathological VAG Signal with Peaks);grid on;% Plot 5: Spectrogram of Filtered Normal Signalsubplot(4, 2, 5);spectrogram(vag_normal_filtered, 256, 200, 512, fs, yaxis);title(Spectrogram of Filtered Normal VAG Signal);xlabel(Time (s));ylabel(Frequency (Hz));colorbar;% Plot 6: Spectrogram of Filtered Pathological Signalsubplot(4, 2, 6);spectrogram(vag_pathology_filtered, 256, 200, 512, fs, yaxis);title(Spectrogram of Filtered Pathological VAG Signal);xlabel(Time (s));ylabel(Frequency (Hz));colorbar;% Plot 7: Normal Signal Analysis Summarysubplot(4, 2, 7);text(0.1, 0.5, {    Normal VAG Signal Analysis Summary:, ...    [Number of Peaks Detected: , num2str(length(peaks_normal))], ...    [Abnormal Detection: , normal_status]}, ...    FontSize, 10);axis off;% Plot 8: Pathological Signal Analysis Summarysubplot(4, 2, 8);text(0.1, 0.5, {    Pathological VAG Signal Analysis Summary:, ...    [Number of Peaks Detected: , num2str(length(peaks_pathology))], ...    [Abnormal Detection: , pathological_status]}, ...    FontSize, 10);axis off;% Display Summary in Command Windowdisp(---- VAG Signal Analysis Summary ----);disp([Normal Signal - Peaks Detected: , num2str(length(peaks_normal)),  -> Abnormal Detection: , normal_status]);disp([Pathological Signal - Peaks Detected: , num2str(length(peaks_pathology)),  -> Abnormal Detection: , pathological_status]);

Normal Signal - Peaks Detected: 72 -> Abnormal Detection: Yes Detection

Pathological Signal - Peaks Detected: 153 -> Abnormal Detection: Yes Detection

图片

​ 本文转载自​​高斯的手稿​​