Cognitive Spectrum Sensing and Detection Using Polyphase DFT Filter Banks

Abstract

In this paper, we present a computationally efficient, signal processing scheme to simultaneously sense and detect multiple active channels for a cognitive radio operating in licensed and unlicensed frequency bands. This new scheme employs a combination of prolate window filter and polyphase DFT filter bank (DFB). A frequency domain power detection scheme is deployed with polyphase DFB to detect multiple active channels based on the sensing output of the DFB. It is shown the detection method achieves comparable performance to the simple power detector with the advantage of low computational complexity leading to reduced delays in cognitive radio networks. The detection performance can be enhanced by exploiting information on the past history of primary user channel access. The theoretical detection threshold based on past information is also derived. The proposed scheme has been analyzed and tested through simulations with parameters based on the IEEE 802.22 standard.