Digital Radio Receiver Responses for Combating Frequency-Selective Fading

Frequency-selective fading in a radio channel, caused by multipath propagation or any other phenomenon, can seriously degrade the effectiveness of digital transmission. Receiver processing that adapts in an appropriate way to the prevailing channel response can strongly reduce the harmful effects of such fading. We examine here the theoretical possibilities of adaptive processing by deriving and analyzing receiver responses for three different criteria. The criteria used permit an analytical approach that is simple and exact. The performance characteristics derived for the three receivers bracket the "best" performance that is possible in practice and reveal possible tradeoffs between performance and practicality. The analysis applies generally to the broad class of quadrature amplitude modulation (QAM) signals, which are assumed to be filtered in the transmitter to enforce spectral emission requirements. For each receiver response we derive a formula for the "power penalty", defined as the increase in transmitter power (over some theoretical minimum) needed to compensate for transmit filtering and channel fading. The power penalty formulas are evaluated for a two-ray multipath fading channel and for each of two common forms of modulation/transmit filtering. Graphical results are given for numerous combinations of the multipath parameters and the transmitter bandwidth-to-symbol rate ratio. The results of this study can be used to predict the probability (time fraction) of multipath outage, provided that the statistics of the channel parameters are known. Some cursory comparisons are made between the new results and those of a previous analysis for fixed receivers. The outage reductions made possible by using appropriate receiver responses are found to be quite large, possibly two or more orders of magnitude.