A fast spectrum sensing for OFDM based on adaptive thresholding

Efficient and accurate spectrum sensing is a necessary part in cognitive radio networks. A simple and efficient spectrum sensing technique is proposed to solve the bad performance of spectrum sensing in non synchronous OFDM signals with low SNR regimes. Through the decomposition of the traditional OFDM signal spectrum sensing technology based on cyclostationarity, the real part and the imaginary part of the correlation function are separated. The two order cyclostation of OFDM signal is analyzed from the real part and imaginary part respectively, and the interference caused by imaginary part correlation function is excluded. According to the estimated noise power and false alarm probability to set the corresponding adaptive threshold to complete the OFDM spectrum sensing by the real part correlation function. The simulation results show that compared with the existing methods the proposed algorithm can effectively reduce the computational complexity and improve the perceptual efficiency, and can realize the fast and accurate of OFDM spectrum sensing.