Overcoming the SNR-wall for energy detection using the autocorrelation of cyclic prefix based OFDM signal

Energy detection is one of the basic schemes of spectrum sensing. In case of energy detection because of the uncertainty in estimating the noise variance there exists a fundamental limit of minimum SNR, called the SNR-wall, below which the signal could not be reliably detected. In this paper, we propose a method to overcome this SNR-wall for OFDM signal by exploiting the autocorrelation property of the cyclic prefix both in AWGN and multipath fading channel. Required SNR for signal detection versus noise uncertainty for a given false alarm probability (0.01) and detection probability (0.99) has been found out in case of both energy detector and the autocorrelation based detector for OFDM signal. Our result shows that there is about 23 dB advantage with our proposed method for AWGN channel and about 12 dB advantage for multipath fading channel over conventional energy detector for noise uncertainty parameter U = 5 with ∼ 5,000 signal samples.