Compressed sensing based blind spectrum sensing in cognitive radio

Cognitive Radios exploit the unutilized spectrum by transmitting opportunistically without disturbing the primary user (PU) or licensed user. So, the sensing should be fast enough that the secondary user (SU) can vacate the spectrum when the PU becomes active. Cyclostationary methods perform better than energy detection methods in low SNR scenarios but they require large number of symbols. The objective of this paper to reduce the number of measurements in blind spectrum sensing (using QR decomposition and backward linear prediction) by employing multiple receive antennas. By using compressed sensing at the receiver front end helps to reduce the complexity of the A/D converter. Furthermore, in this paper, simulation results show that our method achieves a good probability of detection in very low SNR even at 50% measurements which is comparable to probability of detection at full measurements.