Widely Linear System Estimation with Zero Complementary Autocorrelation Sequences

In this paper, the identification of widely linear (WL) systems using sequences with an impulse-like periodic autocorrelation and a zero complementary periodic autocorrelation is addressed. Closed form expressions for sequences with unitary peak to average power ratio (PAPR) suited for the identification of these systems are presented. The analysis shows that the filter impulse responses of the WL system can be estimated from the second order statistics of the system output and the probing sequence. Numerical simulations are provided to verify the variance of the estimation error.