Coded transmission systems for the orthogonal relay channel

(eng) Communication over wireless channels is affected by fading, i.e., fast random variations of the received signal strength. To prevent fading from causing communication outages, a receiver can acquire several copies of the signal and recombine them to average individual variations. In cooperative networks, terminals achieve this by propagating the signal along multiple paths from its source to its destination. The orthogonal relay channel is a simple model for such cooperative transmission, in which a relay helps a source and a destination communicate. Protocols for relaying are separated into three categories, depending on the kind of processing at the relay. In Decode-Forward (DF) protocols, the relay decodes, re-encodes and forwards the signal it receives. In Amplify-Forward (AF) protocols, the relay simply amplifies the signal before forwarding it. And in Compress-Forward (CF) protocols, the relay forwards a compressed version of the signal it observes. This thesis is concerned with the design and analysis of relaying protocols, with a focus on cases where the relay is not able to decode the signal perfectly. A theoretical analysis of the CF protocol shows that the optimal compression method, Wyner-Ziv coding, often reduces to simpler rate-distortion coding when the relay has not enough information on the state of the channels. Practical implementations of relaying protocols follow. The first implementation, Soft Decode-Forward, is a DF relay that forwards the data with an amplitude reflecting its reliability. The second implementation is a receiver for the DF protocol, that takes into account the uncertainty at the relay. The third is a receiver for the CF protocol that jointly decodes the signals from the source and from the relay. Comparisons show that all these implementations outperform usual protocols. In hybrid schemes over wireless channels, the CF protocol with joint reception is the most promising alternative.