On Coding for Fading

Wireless mobile communication systems are characte:ized by time-varying propagation channels, generally rejwed to as 'Ifading chaiinels". Under this conimon name, me find a variety of channel models, suited to diferent app'lications. The rationale behind the selection of coding schmes for wireless systems depends critically on the underlying fading channel model, which in turns depends on the a,oplication. ln this paper, we discuss the coding design criteria for some classical fading channel models. rf the chanrlel is not stationary, as if happens fcJr example in a mobile-i.adio commiinication system where it may fluctuate in tinil? between the extremes of Rayleigh and AWGN, then a code designed to be optiniiim for a fixed channel model plight perform poorly when the channel varies. Therefore, a liptimal code for a given channel may be iicttially suboptirwm for a substantial fraction of time. In these conditions. the issue of robiistness plays an important role: a good code slioiild provide acceptable performance over a wide ran;Te of channel models. A different approach to efficient commtinication over fading channels consists of "clianging the game's riiles" and modifying the propagation channel b,y introducing di'versity andlor exploiting transmitter channel state inforviation. For example, if has been shown that antenna reception diversity witli maximal-ratio combining turns qmptotically a fading channel into an AWGN channel. Then, optimal codes for the latter zuill provide good performiince over wireless channels if used jointly with antenna reception diversity. More recently, join t antenna transrriission diversity and coding has gained a great attention, becriiise of its capability of providing lnrge spectral efficiency. Finally, in some applications the transmitter can be provided 7uith some cliannel state information. Then, it can coiwteract the efect offading by dynamically selecting the appropriate coding scheme andlor by allocating the transmitted power.