Signal extraction and delay estimation by exploiting cyclostationarity

In modern wireless communication systems, the receiver has to operate in the presence of other interfering signals. It has to extract the desired signal and estimate the time delay caused by the channel, prior to the detection of the symbols. In this paper, we present an algorithm that extracts the desired signal in the presence of interferers and estimates the delay of the signal. The algorithm is based on the second order cyclic statistics of the signal. The delay is estimated using the cyclic auto-correlation and we propose a new cost function based on the second order lag product to extract the signal. We then extend the algorithm to the separation of multiple synchronous signals.