Performance analysis of spatially-distributed cooperative networks

• Main Author: Tukmanov, Anvar
• Published: University of Newcastle upon Tyne 2015
• Subjects: 658.4
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.682613

Applications of cooperative communications have attracted considerable attention in academia and industry in the past decade for their potential to exploit network densi cation in meeting the growing demand for data services. However, analytical methods capable of explicitly capturing the impact of the spatial domain on system performance are still rare. The aim of this thesis is to study cooperation between spatially-distributed nodes with the purpose to enhance relevant analytical methods. New approaches to performance analysis of node cooperation and several useful relations are developed in this work in the following three areas. First part of this thesis investigates broadcasting as an important method for TV and network signalling distribution. Cooperative broadcasting (CB) has been generally studied under the assumptions of asymptotically dense or large networks, which rarely hold in practice. In this work, a method to analyse the latency of CB in nite networks is developed using stochastic geometry. New useful relations and inter-node distance distributions are derived, highlighting interesting network characteristics. Second part of this thesis studies relay selection (RS), recognised as a way to reduce overheads arising from cooperative communications. In this thesis, a method for analysing RS is developed based on point processes theory. Presented approach is simpler and more intuitive compared to known methods. This has allowed obtaining exact expressions for outage probability of relay-assisted communication. Additionally, analysis of the sources' contention for relays has revealed that relays can be treated as a scarce resource. Finally, proposed methods are further extended to account for imperfect channel state information (CSI). Practical RS in presence of CSI imperfections remains an active research area, however the aspect of cooperating nodes' spatial distribution remains unexplored. This thesis introduces a novel approach to account for variable levels of CSI accuracy and for the spatial distribution cooperating relays.