Blind identification and equalization of MA channels for QAM communication signals

Presents a new method for blind identification and equalization of moving average (MA) channels for QAM communication signals. Previous results on identifiability of nonminimum-phase MA systems show that the statistics of second-order cyclostationarity at the output are sufficient for identifiability of almost all such systems. Consequently, a new method using only second-order statistics of cyclostationarity in a least-squares framework is proposed. The method can also be interpreted as maximizing the conditional likelihood of the received data under reasonable assumptions.<
>