A Distributed Resource Block Assignment Scheme for Relay-Assisted Cellular Networks With Self-Organizing Terminal Relays

Terminal relaying offers an effective means for improving the performance of OFDMA-based wireless networks. This is because, in addition to offering implementable high orders of spatial diversity to Wireless Terminals (WTs), the number of Terminal Relays (TRs) naturally increases with the number of WTs. However, effective coordination between these TRs becomes a cumbersome task with the increase in the number of TRs. In particular, when their number is large, centralized coordination might not be efficient due to the overhead necessary for Resource Blocks (RBs) coordination. To address this drawback, in this thesis we propose an autonomous resources coordination scheme. In this scheme, the TRs assign RBs to incoming WTs in a way that minimizes the number of hit occurrences at which the same RB is assigned to other WTs. The set of RBs to be assigned by each TR is endowed with a cyclic group structure. This facilitates sequences generation and enables each TR to have access to the entire set of available RBs. Hence, the proposed scheme is particularly beneficial in terminal relaying systems in which the distribution of the WTs is nonuniform and the Channel Quality Indicators (CQIs) are not available. One problem that arises in the proposed scheme is the complexity involved in selecting the best performing cyclic sequences. To address this problem, we propose an analytical approach in which a metric is employed to systematically select the cyclic sequences. This work is further extended by proposing a graphical cyclic sequences selection technique which significantly reduces the complexity of the selection process. We then exploit the cyclically structured assignment sequences to enhance the performance of the proposed scheme. To do so, we propose an algorithm to utilize the hit occurrences in estimating the load level of the system. This information is then exploited in updating the assignment sequences in order to avoid future hits. Simulation results show that the performance of the proposed scheme is significantly better than that of assignment schemes proposed in the literature. ii