An efficient constraint programming approach to signal recovery in compressed sensing

Compressed sensing (CS) allows for the useful information conveyed by a signal to be completely acquired in a few measurements from which the original signal can be accurately reconstructed. This is made possible because of the sparsity property of the original signal, and the existing powerful optimization theory that gave birth to numerous recovery algorithms. With the aim of CS performance improvement, in this paper, we propose the constraint programming (CP) solvers as an alternative to the classical recovery algorithms in the CS process. We show that contrarily to the conventional recovery algorithms, the proposed approach is sensitive to the sensing matrix variance, and provides better performance. Besides, we demonstrate that even non-sparse signals can be recovered with CP-based signal recovery.