Performance of optimum detector for cyclic prefixed OFDM with induced cyclostationarity for spectrum sensing in cognitive radio

Abstract

Cognitive radio offers a solution to spectrum scarcity by intelligently detecting unutilized licensed spectrum. This requires reliable detection of incumbent primary users at sufficiently low SNR. But sensing OFDM signals is particularly challenging as any inherent cyclostationarity is destroyed due to orthogonality of sub-carriers. Induced cyclostationarity can be used to introduce a controlled correlation in the message signal and produce cyclostationary features that can aid spectrum sensing. This paper derives an optimum maximum likelihood test statistic from first principles for detection of a cyclic prefixed OFDM (CP-OFDM) signal with induced cyclostationarity. It adopts a vector matrix model for the CP-OFDM signal in the AWGN scenario and empirically evaluates the detection probability through a Neyman-Pearson test. Simulation results show that a higher probability of detection can be achieved by using the proposed optimum detector for induced cyclostationarity as compared to energy detection and simple cyclic prefix detection. Also the performance scales with number of subcarriers and induced correlation.