An Efficient Filter Banks Based Multicarrier System in Cognitive Radio Networks (invited paper)

Abstract

In cognitive radio techniques, OFDM is usually regarded as the physical layer candidate. However, the weaknesses of the OFDM technique, i.e., using plain FFT for spectral analysis , decreased bandwidth efficiency due to CP (cyclic prefix), high out-of-band emission, have been pointed out and the introduction of filter banks based multicarrier (FBMC) system has been advocated by a number of authors. In this paper, we propose an efficient FBMC system for cognitive radio network. At the transmitter, we propose a decimation transform decomposition method to eliminate the unnecessary zero operations. At the receiver, we utilize the analysis filter banks to sense the spectrum bands. In order to conquer the shortages of the traditional filter banks, we propose a multistage analysis filter banks, which can reduce the computational complexity while improve the detection precision when used to sense the spectrum bands. And with an adaptive threshold scheme in the power estimator, the threshold can be kept very close to the noise power, which can increase the detection probability especially in the condition of low SNR.