摘要

In this paper, we study energy detection-based cooperative Spectrum Sensing (SS) over frequency selective channels in cognitive radio networks. We treat primary signal as an unknown deterministic signal with known power. We assume a generalized fading model for the sensing channel with multipath and Nakagami-m distribution and with a smaller Coherence Bandwidth (CB) than the bandwidth of the primary signal. We investigate an energy combining algorithm in frequency domain to take advantage of the diversity of multiple (relatively flat) sub-bands. We also show that although the channel gains of multiple paths may be uncorrelated, the frequency-domain gains will be correlated. Hence, we take the correlation among these sub-bands into account to improve detection performance. In this frequency domain average likelihood ratio detector (FD-ALRD), the likelihood ratio function of the normalized energies over multiple sub-bands is averaged over the channel distribution in frequency domain. Simulation results show that utilizing the frequency-domain diversity in the FD-ALRD improves the detection performance. Furthermore, exploiting the correlation in the frequency domain results in further improvement.

  • 出版日期2014-6