A novel sparse multiband signal reconstruction method by using Periodic Nonuniform Sampling

In this paper, we present a method of reconstructing multiband signals with arbitrary Spectrum Support Function (SSF). The signal reconstruction method consists in the theorem of Minimum Energy Reconstruction (MER) which is extended for multiband signals. The method allows use of low sampling rates close to Landau rate to achieve the goal. Signal samples are collected by using Periodic Nonuniform Sampling (PNS). The sparse multiband signal reconstruction consists of two stages. In the first stage, the conservative SSF is refined with use of the samples collected by PNS. In the second stage, the MER for non-baseband signals is used to reconstruct the sparse signal spectrum.