Sensing orthogonal frequency division multiplexing systems for cognitive radio with cyclic prefix and pilot tones

The detection of orthogonal frequency division multiplexing (OFDM) for cognitive radio is considered in this paper. A frequency-selective fading channel is considered and the receiving process is modeled with timing and frequency offsets. Firstly, the authors propose a new decision statistic based on time-domain cross-correlation of the cyclic prefix (CP) embedded in OFDM signals. The probability distribution functions (PDFs) of the statistics under both hypotheses of primary signal absence and presence are derived. Estimation of the timing and frequency offset is obtained through the maximum likelihood method and the received signals are modified. Then another new decision statistic based on frequency-domain cross-correlation of the pilot tones (PTs) is proposed whose PDF is also analyzed. Then, through the likelihood ratio test, the authors utilize CP and PT jointly and propose a global test statistic. The theoretical probabilities of false alarm (PFA) and detection are derived, and the theoretical threshold for any given PFA is proposed. The simulation results show that the proposed spectrum-sensing scheme has excellent performance, especially under very low signal-to-noise ratio (SNR).