Noise Cancellation using Normalized LMS and Leaky LMS in MATLAB for Speech enhancemen

hi everybody,
I am doing research in matlab- R(2010'a) for noise cancellation using adaptive filters. Here, i am analysing Normalized LMS, and Leaky LMS algorithms. But I don't know how to create the filter using matlab, and i must doing it with source code, not with simulink. Please could some experts can help me to make the source code for both algorithms.



thanks,
iman

imanrahman said:

hi everybody,
I am doing research in matlab- R(2010'a) for noise cancellation using adaptive filters. Here, i am analysing Normalized LMS, and Leaky LMS algorithms. But I don't know how to create the filter using matlab, and i must doing it with source code, not with simulink. Please could some experts can help me to make the source code for both algorithms.

thanks,
iman

Click to expand...

Hi imanrahman, you can implement your research idea in matlab simulink easily but in terms coding you can find it in matlab demos and costruct your own simple code but unfortunately i don't know how efficiently it works.....

Hope it may be Helpful...

LMS method to findfilter coefficients of order 2. here is code

plot is showing opposite result because of some mistake which i couldn't found so it is diverging instead of convergence.here is code
M=2; %tap weighth length
a(1:M)=[1 2.0260]% 2.1480 -1.1590];
l=M;mu=0.3;tol(1:M,1)=1e-4';

U=rand(1,l)';%input vector of length 10
R=U'*U; % R matrix
Wo=a';
d=Wo'*U;
P=U*d';
W_hat(1:M,1:10000)=0;
% %
i=2;
d_hat(1)=W_hat,1)'*U;
deltaJw,1)=-2*U*d'+2*U*U'*W_hat,1);
while(i<10000)
d_hat(i)=W_hat(1:M,i)'*U;
W_hat,i)=W_hat,i-1)-mu*U*(d-d_hat(i));
deltaJw,i)=-2*U*d'+2*U*U'*W_hat,i);

if (abs(W_hat,i)-Wo))<tol,1)
break
end
i=i+1;
end
figure(2)
plot(W_hat(1,,W_hat(2,,'or')
hold on
plot(Wo(1,1),Wo(2,1),'.')
axis ([-3 3 -3 3])
xlabel('w1','FontSize',10)
ylabel('w2','FontSize',10)
grid on
title('w2 vs w1 with step size 0.3','FontSize',14)

- - - Updated - - -

hi this is my recent work.i will move to NLMS method after gatting good reult for LMS method. if you can debug the problem in above code then i am waiting.