Cooperative Compressed Sensing for Wide-Band Spectrum Detection

We present a cooperative compressed sensing scheme for wide-band spectrum detection in cognitive radio where analog-to-information converter are applied to map a continuous-time signal to a discrete sequence of measurements. The measurements are collected from each second user to a fusion center where a novel cooperative compressed sensing recovery algorithm named SCSMP is utilized to reconstruct the sparse vectors of the wide-band estimated spectrum. The spectrum is then used to make a decision on spectrum occupancy. We compare the performance of this cooperative scheme with single SU scheme. Meanwhile, the performance of SCSMP algorithm and SOMP algorithm is evaluated by MSE and detection time.