Cooperative strategies for Gaussian parallel relay networks

Abstract

In cooperative communication networks, one fundamental and important topic is to understand the performance limits in such kind of networks. That leads to a growing demand on the area of network information theory and especially the topic of relay networks. The parallel relay network, in which two or more relays arrayed in parallel to assist the source to transmit message to the destination, is one kind of important relay network. Furthermore, multi-terminal and bidirectional extensions should be considered if we want to model more general scenarios. Based on the above two considerations, we developed two kinds of networks respectively. One is two-user parallel relay network and the other is two-way two-relay network. To simplify the aspects which we are not interested, we assume that there is no direct link between the source and destination or between the relays, and that the channel state is ¯xed and perfectly known at all nodes. In the two-user parallel relay network, we assume there are two source nodes and two relay nodes. The destination is assumed to be equipped with multiple antennas. This models an uplink transmission to the base station in a mesh network. For this model, an outer bound and coding strategies including hybrid schemes based on amplify-forward (AF), decode-forward (DF) and compress-forward (CF) are derived. For each strategy, the channel conditions under which the capacity can be achieved are derived. Compared with the single user network, the CF approach turns out to be more attractive in a multiuser setting. Apart from the analysis in median signal-to-noise ratio (SNR) region, the performance of di®erent strategies under low SNR regime and high SNR regime is also investigated. The two-way two-relay network is a bidirectional parallel relay network, in which two terminals want to exchange messages with the help of two parallel relays. In our model, we assume each terminal has multiple antennas and each relay has a single antenna. Likewise, an outer bound is given for comparison and several dif- ferent cooperative strategies based on AF and DF are proposed. We evaluate the performance of these strategies under di®erent channel conditions. It is shown that not only the basic strategies have their best performance in some scenarios, but also the mixed strategies match well under certain circumstances. Some strategies can achieve or approach the outer bound under certain conditions. Furthermore, the high SNR and low SNR performance is studied.