# Required number of measrments for signal recovery in a compressed sensing MMV problem?

For multiple measurement vector (MMV) problem
$Y=AX$
where $A$ is $m \times n$ sensing matrix and $X$ is $n \times L$ matrix haveing K non zero rows.

What are the necessary conditions on the number of measrments $m$ and sparsity K to recover the signal $X$ from $Y$ using OMP?

• the question here is: recover with which degree of error? – Marcus Müller Apr 26 '18 at 11:37
• As the degree of error is also dependent on the number of measrements, $m$ should be a function of the degree of error, coherence of the sensing matrics, $n$ and $L$ . – Digi1 Apr 26 '18 at 12:03