Spectrum sensing for OFDM systems employing pilot tones

Abstract

OFDM spectrum sensing algorithms are presented in this paper. The proposed spectrum sensing algorithms are based on Time-Domain Symbol Cross-Correlation (TDSC-MRC and TDSC-NP methods) and can be applied to all existing wireless OFDM systems. The statistical behaviors of the TDSCbased spectrum sensors are explicitly analyzed. The sensing performances of the TDSC-based spectrum sensors in terms of misdetection probability are derived in detail. In addition, the spectrum sensing method employing the Cyclic Prefix of the OFDM modulated signals (CP method) is described for comparison purposes. The DVB-T Standard [5] is adopted as an application example to illustrate the proposed spectrum sensing algorithms. Simulation results show that the TDSC-MRC method outperforms the CP method for all values of CP ratio considered. The TDSC methods have the advantage that the detection performances are the same for different CP ratios, while the detection performance of the CP method degrades dramatically when the CP ratio becomes small. From simulation results, for a sensing time of 50 ms, the TDSC-MRC method can achieve misdetection probabilities of 0.1 and 0.01 with respect to a false alarm probability of 0.01 for SNR values of -20.5 dB and -19.5 dB, respectively.