|M.Sc Thesis||Department of Electrical Engineering|
|Supervisors:||Assoc. Prof. Steinberg Yossef|
|Distinguished Prof. Shamai )Shitz( Shlomo|
The transmission of information over a discrete-time flat fading channel with an average power constraint is considered. The scenario in which the transmitter alone has knowledge of the fading levels, as side information, is assumed. This knowledge may be provided to the transmitter in either a causal or a non-causal manner. Upper and lower bounds on the relevant capacities are derived for each of the two cases using techniques in convex optimization and information theoretic methods. Arimoto-Blahut like iterative algorithms, to numerically compute the capacity in each case when an average power constraint is imposed on the transmission, are presented. The tools developed are then applied to several fading models of interest, specifically, to the On/Off fading channel and to some restricted cases of a Rayleigh fading channel. The new results are compared with known results for other scenarios of side information available to the communication system, such as, the fully informed scenario, when both the receiver and the transmitter have the side information, and with the non-coherent scenario, when neither the receiver nor the transmitter have the side information. Finally, in the Rayleigh fading model the capacity per unit cost is also examined and it is shown that transmitting at an arbitrarily low Eb/N0 will sustain reliable communication at zero spectral efficiency, regardless of the causality/non-causality nature of the available side information, as is the case in the fully informed scenario.