Novel Spectrum Sensing Algorithm for Detecting OFDM Signals

In this paper, a novel spectrum sensing algorithm, which is based on the principle of likelihood ratio detection, is proposed to detect OFDM signals. In the algorithm, a detection statistic is constructed with the correlation of cyclic prefix of OFDM signals. And then, the likelihood functions of the detection statistics in two hypotheses, the primary user is present or absent are approximated with the Gram-Charlier series expansion. Finally, the spectrum sensing decision is made by comparing the likelihood functions. It overcomes the predicament that the decision threshold is difficult to trade off the spectrum detection probability and false alarm probability in traditional autocorrelation detection algorithm. Simulation results show that the algorithm proposed has about 4∼8 dB advantage over the traditional autocorrelation detection algorithm.