Cooperative sequential binary hypothesis testing using cyclostationary features

A cognitive radio should be capable of finding the best spectrum band for communication depending on primary transmissions, the ambient noise level and interference. The first step in achieving this goal is to sense the existence of primary and secondary transmitters in various channels. In addition to the problem of signal detection, there is a need to distinguish between different signals at very low SNR. In this paper, the spectral correlation function is used for hypothesis testing. The sufficient statistic for feature vector based detection in the presence of timing uncertainty is derived. Sequential detection is used to decrease the average number of samples required for testing. Theoretical expressions for the stopping time at low SNR are derived for the AWGN channel. In a fading environment, the performance is evaluated using an approximate expression for the distribution of spectral correlation function. Monte-Carlo simulations verify the accuracy of the theoretical expressions.