Distributed space-time coding for cooperative networks

Cooperative networks exploit the broadcast nature of wireless channels to gain spatial diversity. This dissertation develops two distributed space-time coding schemes for frequency-nonselective channels, and extends and optimizes the previously proposed distributed space-time filtering scheme for frequency-selective channels. Unlike many other distributed space-time coding schemes in the literature, the decentralized schemes proposed in this thesis are designed specifically for wireless networks with a large set of N decode-and-forward relay nodes. At any given time, an a priori unknown subset of nodes acts as relays to cooperatively assist the communication between the source and destination node. To facilitate the cooperation, each physically distributed single--antenna node in the network is assigned a signature vector (signature matrix for multiple-antenna nodes) or signature filter vector. We show that the proposed schemes guarantee a certain diversity gain regardless of which relay nodes in the network cooperate. In addition, the decoding complexity of the various schemes is independent of N and the (random) number of active nodes, and receiver designs advocated originally for traditional co-located antenna systems can be applied. The first scheme is called distributed space-time block coding. Depending on the chosen design, it allows for low-complexity coherent, differential, and noncoherent detection. The second scheme is referred to as distributed space time trellis coding. We show that distributed space-time trellis coding inherits the coding gain of space-time trellis codes over space-time block codes even in the cooperative setting. The two aforementioned schemes are designed for frequency-nonselective fading channels. Finally, we optimize and extend distributed space-time filtering, proposed originally by El Gamal and Aktas, to frequency-selective fading channels. For each scheme, the design criteria are derived. Then, using mathematical tools and classical optimization techniques, efficient methods for the design of the set of signature vectors, signature matrices, and signature filter vectors are provided. Furthermore, the achievable diversity gain, as well as the loss entailed by the distributed implementation of each scheme, are characterized and verified via simulations. Finally, we apply noncoherent distributed space-time block coding to a practical wireless sensor network and show that the proposed scheme is a promising solution for cooperative communication in future sensor, ad hoc, and wireless networks.Applied Science, Faculty ofElectrical and Computer Engineering, Department ofGraduat