Why OMP is iterative? Can we find the support in one step?

1 reply · 1K views · last reply
Awais088
Messages
1
Reaction score
0
In OMP why the 'support' of the received signal is not found in one step, by selecting the $K$ maximum correlated columns of the sensing matrix $A$? where $K$ is the sparsity of the signal. Like in MATLAB

[val ind] = sort(A'*Y); %Y is received signal
signal_Support = ind(1:K); % The result is almost the same as that of normal OMP

Why the support is found in an iterative fashion? As more computation is needed at each iteration.
 
Awais088 said:
In OMP why the 'support' of the received signal is not found in one step, by selecting the $K$ maximum correlated columns of the sensing matrix $A$? where $K$ is the sparsity of the signal. Like in MATLAB

[val ind] = sort(A'*Y); %Y is received signal
signal_Support = ind(1:K); % The result is almost the same as that of normal OMP

Why the support is found in an iterative fashion? As more computation is needed at each iteration.
Can you show us your computations in more detail now? Hopefully you've made some progress on this. See this paper for example:

https://pdfs.semanticscholar.org/425b/1a11936405eb4f8c93dafc33fcc1b485f2a5.pdf
:smile: