Spectrum Sensing Based on Spectral Features of Primary Signals

Spectrum sensing is one of the key technologies for cognitive radios.A spectrum sensing method based on spectral features of primary signals is discussed and a new decision statistic is proposed which is not influenced by noise power uncertainty.The procedure of the algorithms is detailed and simulations are conducted to evaluate the performance of the method using ATSC(Advanced Television Systems Committee) as the primary signal.Results show that using power spectral density of primary signal obtained from larger point of fast Fourier transform(FFT) can result in higher probability of detection.When the point of FFT is kept the same,increasing number of averaging in periodogram calculation helps little to increase the detection performance.Moreover,simulation results under different observation time durations reveal that increasing the observation time can help improve the detection performance of the method.