Dynamic adjustment of sparsity upper bound in wideband compressive spectrum sensing

Compressive sensing (CS) techniques play a key role for fast spectrum sensing in cognitive radio (CR) as it allows perfect signal reconstruction at sub-Nyquist sampling rates. However, for traditional compressing sampling approaches, the sparsity level of a signal is normally assumed as static and known, is impossible in practice. Traditionally, a statistical value of sparsity level upper bound is used as the sparsity level for signal reconstruction. In this paper, we proposed a dynamic adjustment scheme to estimate signal sparsity accurately and recover signals efficiently. In the proposed scheme, a Shrink Algorithm and Enlargement Algorithm are designed to adaptively adjust the value of sparsity level upper bound. Simulation results show that if sparsity level is too large or too small, our proposed scheme can adjust it to an proper value.