Analytical performance bound on a correlation matrix based blind equalizer

Blind equalizers are used to remove the intersymbol interference (ISI) caused by a communication channel without knowing the transmitted sequence. Blind equalizers can be obtained from the second-order cyclostationary statistics of the received sequence. The successful equalization of a mixed-phase communication channel often requires a suitable delay at the output of the equalizer. An arbitrary delay blind equalizer can be obtained with the correlation matrix (CM) method. In this paper, an analytical lower bound for the mean-squared error (MSE) of the CM-equalizer is derived. The MSE is evaluated for example channels and simulations are used to verify the analytical results. For comparison, simulation results obtained with three other algorithms are given. The equalizer performance is studied in simulation as a function of the magnitude of the first channel coefficient.