Equalization on fast flat-fading channels based on signal-to-interference ratio optimization

An optimization criterion for inverse filtering (equalization) of signals transmitted over a Rayleigh flat-fading channel is presented. This criterion is based on the maximization of the ratio of the equalizer output power to the combined power of the intersymbol interference and the noise present at this output. A method to isolate and compute the power of a function of the joint interference signals is presented. This algorithm makes use of the signal innovations concept, in decorrelating a version of the received faded signal to produce the interference estimate. It does not require carrier phase recovery, as do conventional equalization algorithms, and it is relatively insensitive to a multiplicative process imposed on the transmitted signal. This fact allows the algorithm to be very effective in fast flat-fading or severe phase noise conditions. Simulations results are presented to show the performance of this approach.