Blind multiband spectrum signals reconstruction algorithms comparison

This paper investigates sparse sampling techniques applied to downsampling and interference detection for multiband radio frequency (RF) signals. To reconstruct a signal from sparse samples is a compressive sensing problem. This paper compares three different reconstruction algorithms: 1) ℓ₁ minimization; 2) greedy pursuit; and 3) MUltiple SIgnal Classification (MUSIC). We compare the performance of these algorithms and investigate the robustness to noise effects. Characteristics and limitations of each algorithm are discussed.