Deterministic Construction of Partial Fourier Compressed Sensing Matrices Via Cyclic Difference Sets

Abstract

Compressed sensing is a novel technique where one can recover sparse signals from the undersampled measurements. This paper studies a $K \times N$ partial Fourier measurement matrix for compressed sensing which is deterministically constructed via cyclic difference sets (CDS). Precisely, the matrix is constructed by $K$ rows of the $N\times N$ inverse discrete Fourier transform (IDFT) matrix, where each row index is from a $(N, K, \lambda)$ cyclic difference set. The restricted isometry property (RIP) is statistically studied for the deterministic matrix to guarantee the recovery of sparse signals. A computationally efficient reconstruction algorithm is then proposed from the structure of the matrix. Numerical results show that the reconstruction algorithm presents competitive recovery performance with allowable computational complexity.