Data transmission in the presence of channel state information feedback

Licentiatavhandling, 2011

Novel communication techniques such as adaptive resource allocation are useful tools to increase the transmission efficiency and reliability of wireless networks. The main idea behind these techniques is to adapt some transmission parameters, such as the transmission rate and power, based on the instantaneous channel quality. Therefore, instead of relying on statistical information, we should estimate the channel coefficients directly at the receivers and somehow inform the transmitters about the obtained complex fading coecients. The more information available at the transmitters, the higher rates can be achieved, with the rate upper bounded by full knowledge channel capacity. However, due to some issues such as the signaling overhead caused by reporting the channel information or the other users interference limits, assuming perfect channel information at the transmitters is an overly optimistic assumption which does not match with reality. This is the main motivation for the present limited channel state information (CSI) feedback systems such as UMTS/WCDMA and for this thesis as well. The thesis aims to study the performance of communication networks utilizing partial CSI feedback. Two standard quantized CSI and automatic repeat request (ARQ) approaches are implemented as tools providing the transmitter with partial channel quality information. The effect of these methods is evaluated under dierent forward channel fading conditions and feedback channel noise levels. While we select the channel average rate as our main performance evaluation yardstick, the results are obtained under different system quality-of-service requirements. Considering a typical single-user communication setup, we investigate the effect of ARQ and quantized CSI approaches under both peak and average transmiission power constraints. Then, the channel average rate is obtained in the case where both quantized CSI and ARQ feedback information are available at the transmitter. Further, considering noisy and noise-free feedback channels, we investigate the effect of an outage probability constraint on the system achievable rates. Finally, we will evaluate the effect of adaptive quantizers and delayed quantized CSI feedback on the performance of the correlated single-user communication setups. The second part of the thesis focuses on the more general spectrum sharing networks where the users are not only affected by the other users interferences but also should consider their quality-of-service requirements. Assuming different levels of channel quality information, the system performance is studied under different transmission power and other users received interference power constraints. Finally, assuming different power constraints, we demonstrate the effect of other quality-of-service requirements, such as outage probability constraints, on the channel average rate. Theoretical and simulation results show that, there is considerable potential for improving the data transmission efficiency of communication networks via channel quality information feedback.