Unified analysis of communication receivers using a Bayesian approach

(eng) A digital communication system is defined mainly by three entities, namely, transmitter, communication medium and receiver. Particular descriptions of the transmitter and the communication medium lead to particular communication scenarios. And the receiver strictly relies on these descriptions to operate. More explicitly it has to know or estimate the important parameters concerning the transmitter and the communication medium, and only then it can retrieve the transmitted information within acceptable error limits. The purpose of this thesis is to introduce a unified analysis of the digital communication receivers which helps to better assess their underlying structure and provide means for alternative designs. This is done by investigating the digital communication systems as a whole network of Bayesian relations, and tracking the belief propagations within the receiver. Such analysis reveals a novel approach to the receiver design problem; where instead of optimizing a given criteria, it has been aimed to identify particular approximations to the optimal receiver which leads to which leads to certain receiver technique. As a consequence, the thesis intends to show that any given receiver structure for any given transmission scheme can be derived with the help of this analysis basically as a simplification of the optimal structure using some approximations. In the thesis two major benefits of the unified analysis are underlined: 1) different receiver techniques that are proposed in the literature may be compared with respect to their implicit approximations. And 2) with the help of proposed approach to the design problem, alternative receiver structures can be derived leading to improved detection performance.