使用FREEEMG系统和MVN系统的SDK获得的EMG和惯性数据的matlab代码.zip

clear load("DataSets\clustering_data.mat"); load("DataSets\withExoEMG.mat"); load("DataSets\withoutExoEMG.mat"); % load("DataSets\maxEMGVals.mat"); fr = 1000; nq_freq = fr/2; cutoff_low = 10; cutoff_high = 150; max_emg_vals = struct; [b,a] = butter(4,[cutoff_low,cutoff_high]/nq_freq,"bandpass"); exp_names = emg_without_exo_data.keys(); channels_muscles=["Right deltoid", "Right Biceps", "Left Deltoid", "Left Biceps"]; for i=1:length(exp_names) exp_names{i} emg_exp_data = emg_without_exo_data(exp_names{i}); clust_data = clustering_without_exo(exp_names{i}); valid_indexes = find((clust_data.ids==2)); recorted_data = struct; channels_data = emg_exp_data.ch_data; WinLen = 4000; % Window Length For RMS Calculation index=1; step=400; fieldnames_emg = fieldnames(channels_data); for j=1:length(fieldnames_emg) ch_name = fieldnames_emg{j}; ch_data = channels_data.(ch_name); figure(2*i -1) subplot(4,1,j) plot(ch_data(1,:)*0.001, ch_data(2,:)) ylim([-1.6e-3, 1.6e-3]); xlabel("time (s)") ylabel("Sensor signal (V)") title(strcat("Sensor output ",channels_muscles(j))) f = fr*(0:length(ch_data(2,:))/2)/length(ch_data(2,:)); four = fft(ch_data(2,:)); four = abs(four/length(four)); four = four(1:(length(four)/2+1)); four(2:end-1) = 2*four(2:end-1); figure(2*i) subplot(4,1,j) plot(f, four) xlabel("Frequency Hz") ylabel("|fft|") title(strcat("Fourier transform ",channels_muscles(j)," data")) ch_data(2,:) = filtfilt(b,a, ch_data(2,:)); if (isfield(max_emg_vals, ch_name)==false) max_emg_vals.(ch_name) = 0; end max_emg_vals.(ch_name) = max(max_emg_vals.(ch_name), max(ch_data(2,:))); % ch_data(2,:) = ch_data(2,:)./max_emg_vals.(ch_name); emg_indexes = round(ch_data(1,:)/10); recorted_data.(ch_name).recorted = ch_data(2, ismember(emg_indexes, valid_indexes)); end emg_without_exo_data(exp_names{i}).recorted = recorted_data; end exp_names = emg_with_exo_data.keys(); for i=1:length(exp_names) exp_data = emg_with_exo_data(exp_names{i}); clust_data = clustering_with_exo(exp_names{i}); valid_indexes = find((clust_data.ids==2)); recorted_data = struct; channels_data = emg_exp_data.ch_data; WinLen = 4000; % Window Length For RMS Calculation index=1; step=400; fieldnames_emg = fieldnames(channels_data); figure for j=1:length(fieldnames_emg) ch_name = fieldnames_emg{j}; ch_data = channels_data.(ch_name); figure(2*(i+3) -1) subplot(4,1,j) plot(ch_data(1,:)*0.001, ch_data(2,:)) ylim([-1.6e-3, 1.6e-3]); xlabel("time (s)") ylabel("Sensor signal (V)") title(strcat("Sensor output ",channels_muscles(j))) f = fr*(0:length(ch_data(2,:))/2)/length(ch_data(2,:)); four = fft(ch_data(2,:)); four = abs(four/length(four)); four = four(1:(length(four)/2+1)); four(2:end-1) = 2*four(2:end-1); figure(2*(i+3)) subplot(4,1,j) plot(f, four) xlabel("Frequency Hz") ylabel("|fft|") title(strcat("Fourier transform ",channels_muscles(j)," data")) ch_data(2,:) = filtfilt(b,a, ch_data(2,:)); if (isfield(max_emg_vals, ch_name)==false) max_emg_vals.(ch_name) = 0; end max_emg_vals.(ch_name) = max(max_emg_vals.(ch_name), max(ch_data(2,:))); % ch_data(2,:) = ch_data(2,:)./max_emg_vals.(ch_name); emg_indexes = round(ch_data(1,:)/10); recorted_data.(ch_name).recorted = ch_data(2, ismember(emg_indexes, valid_indexes)); end emg_with_exo_data(exp_names{i}).recorted = recorted_data; end %% exp_names = emg_without_exo_data.keys(); for i=1:length(exp_names) exp_names{i} emg_exp_data = emg_without_exo_data(exp_names{i}); clust_data = clustering_without_exo(exp_names{i}); valid_indexes = find((clust_data.ids==2)); recorted_data = struct; channels_data = emg_exp_data.ch_data; WinLen = 4000; % Window Length For RMS Calculation index=1; step=400; fieldnames_emg = fieldnames(channels_data); for j=1:length(fieldnames_emg) ch_name = fieldnames_emg{j}; ch_data = channels_data.(ch_name); ch_data(2,:) = filtfilt(b,a, ch_data(2,:)); ch_data(2,:) = ch_data(2,:)./max_emg_vals.(ch_name); emg_indexes = round(ch_data(1,:)/10); recorted_data.(ch_name).recorted = ch_data(2, ismember(emg_indexes, valid_indexes)); window_indexes = 1:step:length(recorted_data.(ch_name).recorted)-WinLen; rmsv = zeros(1, length(window_indexes)); arvv = zeros(1, length(window_indexes)); md_freq = zeros(1, length(window_indexes)); mean_freq = zeros(1, length(window_indexes)); index=1; for k=window_indexes signal_cutted = recorted_data.(ch_name).recorted(k:k+WinLen); rmsv(index) = rms(signal_cutted,2); arvv(index) = sum(abs(signal_cutted),2)/WinLen; md_freq(index) = medfreq(signal_cutted,fr); mean_freq(index) = meanfreq(signal_cutted,fr); index=index+1; end recorted_data.(ch_name).rmsv = rmsv; recorted_data.(ch_name).arvv = arvv; recorted_data.(ch_name).md_freq = md_freq; recorted_data.(ch_name).mean_freq = mean_freq; figure(15) subplot(4,1,j) hold on plot(rmsv) ylim([0.05, 0.22]) figure(16) subplot(4,1,j) hold on plot(arvv) ylim([0.05, 0.22]) figure(17) subplot(4,1,j) hold on plot(md_freq) ylim([40, 65]) figure(18) subplot(4,1,j) hold on plot(mean_freq) ylim([40, 65]) end emg_without_exo_data(exp_names{i}).recorted = recorted_data; end exp_names = emg_with_exo_data.keys(); for i=1:length(exp_names) exp_data = emg_with_exo_data(exp_names{i}); clust_data = clustering_with_exo(exp_names{i}); valid_indexes = find((clust_data.ids==2)); recorted_data = struct; channels_data = emg_exp_data.ch_data; WinLen = 4000; % Window Length For RMS Calculation index=1; step=400; fieldnames_emg = fieldnames(channels_data); for j=1:length(fieldnames_emg) ch_name = fieldnames_emg{j}; ch_data = channels_data.(ch_name); ch_data(2,:) = filtfilt(b,a, ch_data(2,:)); ch_data(2,:) = ch_data(2,:)./max_emg_vals.(ch_name); emg_indexes = round(ch_data(1,:)/10); recorted_data.(ch_name).recorted = ch_data(2, ismember(emg_indexes, valid_indexes)); window_indexes = 1:step:length(recorted_data.(ch_name).recorted)-WinLen; rmsv = zeros(1, length(window_indexes)); arvv = zeros(1, length(window_indexes)); md_freq = zeros(1, length(window_indexes)); mean_freq = zeros(1, length(window_indexes)); index=1; for k=window_indexes signal_cutted = recorted_data.(ch_name).recorted(k:k+WinLen); rmsv(index) = rms(signal_cutted,2); arvv(index) = sum(abs(signal_cutted),2)/WinLen; md_freq(index) = medfreq(signal_cutted,fr); mean_freq(index) = meanfreq(signal_cutted,fr); index=index+1; end recorted_data.(ch_name).rmsv = rmsv(1:750); recorted_data.(ch_name).arvv = arvv(1:750); recorted_data.(ch_name).md_freq = md_freq(1:750); recorted_data.(ch_name).mean_freq = mean_freq(1:750); figure(19) subplot(4,1,j) hold on plot(rmsv(1:850)) ylim([0.