Blind multi-path diversity for robust primary user detection in cognitive radio

The authors prove that multi-path diversity improves the detection performance of primary users for cognitive radio users. They analyse the relationship between the performance and the diversity order. The result shows that robust performance gain is hard to achieve by choosing fixed diversity order in unknown and varying channel states. Therefore they propose here a blind multi-path diversity auto-correlation ratio detection scheme that can enable dynamic diversity order to adapt to the channel state. Further, they extend this scheme to multi-antenna scenario where the cross-correlation information of the received signal at each pair of antennas is exploited. Simulations show that the proposed schemes perform as well as or even better than the ideal energy detection. One can use these schemes to detect all types of wireless signal without having the knowledge of the targeted signal, the channel and the noise power. Besides, unlike in other robust detection schemes, the thresholds are asymptotically independent of the sample size, which facilitates the flexible sensing period adjustment.