A classification of OFDM signals with or without DFT precoding based on high-order moment

In order to assist dynamic spectrum access in the cognitive radio framework, a precise real-time identification of the signals sent by either the primary or secondary users should be necessary. Orthogonal frequency-division multiplexing (OFDM) signals with and without DFT precoding are commonly adopted in the recent wireless communications standards. Therefore, the accurate classification of such candidate signals would help detecting whether the user occupying the channel is primary or secondary, under the assumption that they use different signal formats. This paper proposes a symbol-by-symbol detector with significantly low complexity and low latency based on the use of high-order moments of their baseband signals. By making use of both time and frequency domain signal representations, the effectiveness of the proposed approach is demonstrated.