Transceiver Beamforming Design and Power Optimization for Cooperative Multi-Source Multi-Destination Networks
MetadataShow full item record
Multi-source multi-destination (MSMD) wireless model is developed for wireless devices to share the limited spectrum resource in the same communication channel, which induces the mutual interference with each other. To reduce or even cancel the mutual interference in MSMD network, various techniques were proposed in literatures, including beamforming design for multi-user multi-input multi-output (MIMO) channels and cooperative relaying strategy. In addition, power control is a critical issue as the performance at each destination depends not only on power allocation of its corresponding source, but also on the power allocation of other sources. An optimal power control strategy helps to manage the interference between each pair of source-destination, thus further improves the overall system performance. Hence a design to optimize power allocation to efficiently apply the MIMO techniques and cooperative relaying protocol becomes more attractive for MSMD networks. In this dissertation, we analyze the performance improvement and develop optimal power allocation strategy for MSMD MIMO and relaying network based on optimize methodology, approximation theory and game theory. We aim to develop optimal strategies with low complexity for power allocation in MSMD network by considering two optimization models. (1) Power minimization subject to Quality of Service (QoS) constraints, including Signal to Noise plus Interference Ratio (SINR) and outage probability. (2) QoS maximization subject to power constraints. First, we develop an optimal transceiver beamforming strategy to minimize the total power of sources subject to SINR constraints for a MSMD network where MIMO technique is applied in both sources and destination. We propose a distributed algorithm to find a Nash optimal solution and provide a necessary condition and a sufficient condition to the existence and the uniqueness of Nash optimal solution. We show that the probability of convergence of the proposed algorithm is higher than earlier work. Though Nash optimal solution is easy to obtain in practical, yet it is not efficient in terms of overall power consumption. Thus we further develop a more efficient algorithm for transceiver beamforming design for MSMD MIMO network with pricing consideration in order to find a Pareto optimal solution, which is more efficient than Nash solution in terms of overall power efficiency. We also compare the proposed algorithm over another two alternatives to show our advantage. In convergence analysis, we provide a sufficient and a necessary condition to the convergence of the algorithm. Different from most of earlier related work, where only single antenna at destination was considered, our transceiver design is also applicable to the scenario where multi-antennas are deployed in each destination. MIMO beamforming technology does open an avenue for us to improve the performance of MSMD network. However, due to space and cost limitation, MIMO technique might not be available in some applications. As an alternative, cooperative relaying is a promising technologies to improve the performance of MSMD network. To discuss the optimal power allocation strategies in MSMD relaying network, we start from the optimal power allocation design in a single-source single-destination relay network with Differential Amplify-and-Forward (DAF) protocol. With our simple approximation of outage probability for a DAF scheme, we are able to propose a simple solution to optimize the power allocation, which is a tight approximation to the accurate optimal power allocation verified by our numerical results. In addition, we also provide an optimal relay location design and joint power and relay location strategy for a DAF relaying protocol to further improve the overall system performance. Differing from the single-source single-destination relay network, the performance analysis for MSMD relay network is more complicated due to the effect of the generated interference between each pair of users. By considering the SINR and its approximation of each user, we are able to develop an optimal power allocation strategy for MSMD relaying network to guarantee the minimal SINR of each user is satisfied. With an approximation of SINR representation, we propose low complexity algorithms to optimize the power assignment for MSMD relaying network, where two scenarios are investigated, minimization of total power subject to SINR constraints and maximization of minimal SINR subject to total power constraint. The results obtained by the proposed algorithms are shown to be very close to those obtained by exhaustive search, while the former outperforms the latter in terms of very low computation complexity.