Abstract
The topic of this thesis is the transmission of a memoryless source over a slow fading channel. It should be noted that when a channel is specified as a slow fading channel, it does not specify whether the channel is flat or frequency selective in nature. The signal transmitted through such a channel is not only disturbed by additive white Gaussian noise but also suffers random amplitude attenuation and phase shift (fading). Several well-known techniques for overcoming the degradation imposed by fading channels are: 1. Channel Coding channel coding combined with bit interleaving, which introduces delay and that may not be acceptable in certain applications. 2. Dual Channels the use of two channels such that the core information is sent over the one with the lowest probability of loss.3. Diversity diversity systems in which replicas of the same information are sent over independent fading channels and combined at the receiver to minimize the probability of error, the maximum ratio combining (MRC) being a perfect example of such systems pendent fading channels and combined at the receiver to minimize the probability of error, the maximum ratio combining (MRC) being a perfect example of such systems. However, in dual diversity systems, if no error occurs in the transmission, half of the transmitted information is of no use, hence a great portion of bandwidth is wasted to protect the signal from worst-case channel conditions. In this thesis, we will be using multiple description source coding which tends to minimize the previous waste of bandwidth by transmitting complementary information over independent channels. Previous works using this approach did not consider a realistic fading (both amplitude and phase) or implement a realistic receiver since the fading coefficients were assumed to be known. In this thesis, we derive and implement a maximum-likelihood (ML) receiver using the Expectation-Maximization (EM) algorithm. The performance achieved using the ML receiver is seen to be comparable to those of previous works, without however making any simplistic assumptions.
Pegnyemb, Telesphore Bertrand (1999). Multiple description source coding for mobile radio channels. Master's thesis, Texas A&M University. Available electronically from
https : / /hdl .handle .net /1969 .1 /ETD -TAMU -1999 -THESIS -P44.