RBF.rar_RBF 神经网络_RBF神经网络 VC_rbf VISUAL_rbf c++

clc; clear; load p2; %Set the variables. Type s = p2; L_s = length(s); dt = (1:L_s)/512; %perform a level 5 decomposition of the signal (again using the db8 wavelet) [C,L] = wavedec(s,5,'db8'); %To reconstruct the level 5 approximation from C A5 = wrcoef('a',C,L,'db8',5); %To reconstruct the details at levels 1, 2, and 3, from C D1 = wrcoef('d',C,L,'db8',1); D2 = wrcoef('d',C,L,'db8',2); D3 = wrcoef('d',C,L,'db8',3); D4 = wrcoef('d',C,L,'db8',4); D5 = wrcoef('d',C,L,'db8',5); %To display the results of the level 3 decomposition figure(1); subplot(2,3,1); plot(A5); title('Approximation A5') subplot(2,3,2); plot(D1); title('Detail D1') subplot(2,3,3); plot(D2); title('Detail D2') subplot(2,3,4); plot(D3); title('Detail D3') subplot(2,3,5); plot(D4); title('Detail D4') subplot(2,3,6); plot(D5); title('Detail D5') %to reconstruct the original signal from the wavelet decomposition structure A0 = waverec(C,L,'db8'); err = s-A0; %To compare the approximation to the original signal figure(2); %subplot(2,1,1);plot(s);title('Original'); %axis off %subplot(2,1,2);plot(A5);title('Level 5 Approximation');%axis off plot(dt,A5); %title('A5'); %axis off xlabel('t /ms'); ylabel('A5 /MPa'); %subplot(3,1,3);plot(err);title('error'); %Remove noise by thresholding. Let's look again at the details of our level 3 analysis. %To display the details D1, D2, and D3, type figure(3); subplot(5,1,1); plot(dt,D1); title('D1'); axis off subplot(5,1,2); plot(dt,D2); title('D2'); axis off subplot(5,1,3); plot(dt,D3); title('D3'); axis off subplot(5,1,4); plot(dt,D4); title('D4'); axis off subplot(5,1,5); plot(dt,D5); title('D5'); axis off %correlation 计算自相关函数 c0 = xcorr(A5); c1 = xcorr(D1); c2 = xcorr(D2); c3 = xcorr(D3); c4 = xcorr(D4); c5 = xcorr(D5); figure(4); dt2 = (-L_s+1 : L_s-1)/512; subplot(2,3,1); plot(dt2,c0); xlabel('t /ms'); ylabel('Correlation A5'); %title('Corr A5') subplot(2,3,2); plot(dt2,c1); xlabel('t /ms'); ylabel('Correlation D1'); %title('Corr D1') subplot(2,3,3); plot(dt2,c2); xlabel('t /ms'); ylabel('Correlation D2'); %title('Corr D2') subplot(2,3,4); plot(dt2,c3); xlabel('t /ms'); ylabel('Correlation D3'); %title('Corr D3') subplot(2,3,5); plot(dt2,c4); xlabel('t /ms'); ylabel('Correlation D4'); %title('Corr D4') subplot(2,3,6); plot(dt2,c5); ylabel('Correlation D5'); xlabel('t /ms'); %title('Corr D5') %period signal detect figure(5); plot(c4); title('Corr D5') %FFT y0 = fft(A5); Py0 = y0.*conj(y0) / L_s; y1 = fft(D1); Py1 = y1.*conj(y1) / L_s; y2 = fft(D2); Py2 = y2.*conj(y2) / L_s; y3 = fft(D3); Py3 = y3.*conj(y3) / L_s; y4 = fft(D4); Py4 = y4.*conj(y4) / L_s; y5 = fft(D5); Py5 = y5.*conj(y5) / L_s; df = 512*(0:L_s/2-1)/L_s; figure(6); subplot(2,3,1); plot(df,Py0(1:L_s/2)); %title('FFT A5') xlabel('\omega /kHz') ylabel('A5功率谱'); subplot(2,3,2); plot(df,Py1(1:L_s/2)); xlabel('\omega /kHz') ylabel('D1功率谱');%title('FFT D1') subplot(2,3,3); plot(df,Py2(1:L_s/2)); xlabel('\omega /kHz') ylabel('D2功率谱');%title('FFT D2') subplot(2,3,4); plot(df,Py3(1:L_s/2)); xlabel('\omega /kHz') ylabel('D3功率谱');%title('FFT D3') subplot(2,3,5); plot(df,Py4(1:L_s/2)); xlabel('\omega /kHz') ylabel('D4功率谱');%title('FFT D4') subplot(2,3,6); plot(df,Py5(1:L_s/2)); xlabel('\omega /kHz') ylabel('D5功率谱');%title('FFT D5') figure(7); plot(df,Py4(1:L_s/2)); figure(8); plot(df,Py5(1:L_s/2)); % perform the actual de-noising, type [thr,sorh,keepapp] = ddencmp('den','wv',s); clean = wdencmp('gbl',C,L,'db8',5,thr,sorh,keepapp); %To display both the original and de-noised signals, type figure(9); subplot(2,1,1); plot(s); title('Original') subplot(2,1,2); plot(clean); title('De-noised'); %differentiate for i = 1; L_s t(i) = i; end deriv1 = diff(A5); erro = s - A5; figure(10); %subplot(2,1,1); plot(deriv1); title('dP/dt') %subplot(2,1,2); plot(erro); title('error'); plot(dt(1:L_s-1),deriv1); %title('dP/dt'); xlabel('t /ms'); ylabel('dP/dt /(MPa/ms)'); %fft of original signal y0 = fft(s); Py0 = y0.*conj(y0) / L_s; plot(abs(y0)); xlabel('\omega /kHz') ylabel('S 功率谱');