Cooperative Distributed Multiuser MMSE Relaying in Wireless Ad-Hoc Networks

We consider a wireless ad-hoc network with single antenna nodes under a two-hop traffic pattern. Two system architectures are investigated in this paper: Either linear amplify-and-forward relays (LinRel) or a distributed antenna system with linear processing (LDAS) serve as repeaters. The gain factors of the repeaters are assigned such that the mean squared error (MSE) of the signal at the destinations is minimised (multiuser MMSE relaying). A scalar multiplier γm ∈ C at each destination m allows for received signals that are scaled and rotated versions of the transmitted symbols. We distinguish two cases: 1) The factors are equal for all destinations: γm = γ. 2) An individual factor γm is chosen for each destination m. Multiuser MMSE relaying essentially realizes a distributed spatial multiplexing gain with single antenna nodes as all source/destination pairs can communicate concurrently over the same physical channel. The main contribution of this paper is the derivation of the MMSE gain factors. We evaluate the relaying scheme in comparison to multiuser zero forcing (ZF) (1) and show that it can outperform the latter in terms of average sum rate and diversity gain. Keywords - cooperative relaying, ad-hoc networks, distributed spatial multiplexing, minimum mean squared error (MMSE)