| In this thesis, two problems in cooperative relay networks are tackled, namely, diversity combining of signals with different modulation levels, and Constellation Rearrangement (CoRe).;The problem of CoRe is defined as finding good bit to symbol mapping for each transmitting node, without changing the modulation level. Through exhaustive numerical search, we propose a good CoRe scheme. Unlike most of the existing CoRe schemes, the proposed CoRe scheme does not use Gray-coding constellation in any of the transmitting nodes. In the context of fixed relays, the proposed CoRe scheme achieves significant gain compared to the conventional scheme and it outperforms the existing CoRe techniques. In the context of nomadic relays, we observe that the proposed CoRe, compared to conventional and other existing CoRe schemes, is the most sensitive scheme to error propagation. More importantly, in order for any CoRe scheme to have better performance than the conventional scheme, the average SNR in the source-relay link must be greater than a threshold value that is a function of both the average SNRs in the source-destination and relay-destination links. Otherwise, the CoRe schemes degrade the BER performance as they amplify the undesirable effect of error propagation. In nomadic relay networks, the proposed CoRe outperforms the conventional and the existing CoRe schemes if and only if the average SNR in the source-relay link is greater than this threshold.;In digital cooperative relaying, signals from the source-destination and relay-destination links are combined at the destination to achieve spatial diversity. The vast majority of the research done in cooperative relaying assumes the modulation level used by both the source and relay to be the same. This assumption does not necessarily hold in next generation wireless networks where adaptive modulation is implemented and in such a case, conventional maximal ratio combining does not work. This raises the need to investigate different receiver structures and understands the performance gain as well as the complexity associated with each receiver. Consequently, performance analysis as well as simulation results of the BER of different receiver structures are presented. We show that performing soft-bit maximal ratio combining is the most attractive receiver structure, as it yields near optimal BER performance with low complexity. In the case of nomadic relay networks, where error propagation is a limiting factor in the BER performance, link adaptive regeneration (LAR) can be effectively used. Moreover, we propose a modified LAR scheme that significantly outperforms its conventional counterpart, without any increase in the complexity. The gain of the proposed scheme is illustrated through both analysis and simulation. |
