Hi !! I hope somebody can help !! N = 8000; % Number of Samples n = 0:N-1; sigmax = 0.01 L = 12 f1 = 101.7 f2 = 142.4 f3 = 231.5 fs = 1000 K = 20 % Number of realizations cumensq = zeros(K,N) en1 = zeros(K,N) % length of filter coefficients wn = zeros(1,L); a = 0.0001; muhat = 1.00 wmat = zeros(N+1,L); % 8001 filters wmat(1,:)=wn; for kk = 1: K % ensemble process xn = cos(2*pi*f1/fs*n)+cos(2*pi*f2/fs*n)+cos(2*pi*f3/fs*n)+sigmax*randn(1,N); xn1 = cos(2*pi*f1/fs*n)+cos(2*pi*f2/fs*n)+cos(2*pi*f3/fs*n); dn = sin(2*pi*f2/fs*n); prepad = zeros(1,L-1) xnlong = [prepad ,xn]; mu = 0.01; en = zeros(1,N); % Matrix : Each Row represents one realization : Column is samples for one Realization ensq = zeros(1,N) for k = 1:N % sample process xx= xnlong(L+k-1:-1:k); yn(k) = xx*wn'; en(k)=dn(k)-yn(k); wn = wn + mu * en(k)*xx %wn = wn + muhat/(a+xx*xx')* en(k)*xx %wn = wn + mu*en(k).*xx; % filter update wmat(k+1,:) = wn; ensq(k)=en(k).*en(k); end en1(kk,:)=en cumensq(kk,:) =ensq; end ensemble = sum(cumensq)/K figure aut=CalcCovarianceMatrix(en1,1,8000) stem(aut) I have trying to get the autocorrelation of the error signal !! As you could see that sigmax is defined above and the square of sigmax should give me the variance and the highest peak value on the autocorrelation plot !! I believe the xcorr is not working according to my standard so i tried this CalcCovarianceMatrix() which is defined as function R = CalcCovarianceMatrix(SignalVector, timeStart, timeEnd) N = length(SignalVector(:,1)); % Initialize R. R_n = zeros(N,N); % Compute the time averaged auto-correlation matrix. for TimeIndex = timeStart:timeEnd, Value = SignalVector(:,TimeIndex)*SignalVector(:,TimeIndex)'; R_n = R_n + Value; end; % Scale the Result. R = R_n/(timeEnd - timeStart + 1); however i still not able to get this peak value of sigmax squared (0.01^2) kindly somebody assist me !! what the helly pelly is the problem thanks Muhammad Ali

# Plot of Autocorrelation Error using LMS in Matlab

Started by ●December 11, 2006