Asymptotically optimal blind equalizers based on cyclostationary statistics

We consider the blind estimation of digital communication channels based on second-order cyclic statistics. Specifically, we derive an asymptotically best consistent estimator of the channel impulse response based an a weighted quadratic-matching criterion between the sample cyclic correlations of the data and those of the FIR channel model. Contrary to existing methods, this approach is optimal in the sense of minimizing the asymptotic variance of the channel estimator. We also present expressions for the asymptotic covariance of the cyclic correlation matching approach which can serve as a benchmark for comparison of different second- as well as higher-order approaches