OFDM Signal Sensing over Doubly-Selective Fading Channels

Cyclostationarity based sensing methods are appealing for OFDM signals under low SNR areas. However, such methods are sensitive to doubly-selective fading channels. In this paper we develop a new sensing method by exploiting the cyclostationarity incurred by the spread of autocorrelation due to the Doppler effect in doubly-selective fading channels. First, we model the doubly-selective fading channel with basis expansion model (BEM). Then we analyze the relationship between cyclostationary statistics of transmitted and received OFDM signals, hence derive the cyclostationary statistics of the received signal. Based on the new cyclostationary signatures of the received OFDM signal spread by the Doppler effect, more suitable cyclic frequencies can be chosen to detect cyclostationary features when signal experiences doubly-selective fading. Simulation results demonstrate that the proposed sensing method provides substantial improvement on detection performance.