In the first part of the thesis, we focus on the design of a complete satellite communication system adopting adaptive beamforming with mobile satellite terminals. Adaptive beamforming can significantly improve the capacity of the system in terms of served STs and power efficiency. For the design of adaptive beamforming system, channel state information (CSI) is critical. Since in a satellite system the propagation delay is too long compared to the coherence time of the channel, the instantaneous CSI is already stale when received for design of a beamformer. However, some part of the channel, the directivity vectors, change sufficiently slowly. We use this partial knowledge of CSI to design a system based on adaptive beamforming.
In order to estimate the directivity vectors, we propose an algorithm based on a least square error criterion. Based on the estimation of directivity vectors, we propose two heuristics approaches for the beamforming design. Additionally, we propose two approaches based on the directivity estimation for the detection of transmitting terminals and possible resolution of collisions in the random access channel of the satellite system. Since the SDMA system performance depends strongly on the spatial locations of the co-existing terminals, we propose low complexity algorithms for frequency allocation. Finally, we simulate the complete satellite system, including a random access channel and a connection-oriented channel. We analyze the system performance and compare it to systems based on conventional fixed beamforming.
In the second part, we consider a block fading interfering channel with two transmission pairs. We assume that both transmitters have perfect knowledge of direct links but have only statistical knowledge of the interfering links. We study the problem of transmission rate and power allocation in an autonomous and decentralized manner in absence of perfect CSI. Resource allocation algorithms based on Bayesian game and optimization are proposed.
