Data Transmission in the Presence of Limited Channel State Information Feedback

Foreword Modern wireless systems include key technologies such as multiuser diversity, adaptive transmission, multiple-input multiple-output (MIMO) processing and cooperative data communication networks. Common for all these technologies is that the feedback of channel state information (CSI) is necessary, as it is exploited for adapting the transmission parameters, scheduling, etc. The CSI is normally obtained by direct estimation of the channel coefficients at the receiver and somehow informing the transmitter about the obtained complex fading coefficients. The more information available at the transmitter, the better performance can be achieved. However, depending on the channel characteristics, number of transmit/receive antennas, etc., the rate of the feedback information may be so high that it will consume much of the capacity in the reverse link, making the whole system impractical. Therefore, it has often been questioned whether the improved system performance due to CSI feedback is worth the additional feedback rate and the increased implementation complexity. Thus, different CSI feedback compression techniques have been of interest during the last two decades. In the thesis, we consider two different approaches, namely, quantized CSI feedback and automatic repeat request (ARQ), providing the imperfect channel quality information at the transmitter. Implementing a quantized CSI feedback scheme, the receiver provides the transmitter with some rough measure of the channel gain before transmission, and the transmitter adjusts its transmission parameters according to this imperfect information. Rough CSI is normally produced by channel gain quantization at the receiver where the set of all possible channel gains is partitioned into a number of non-overlapping regions. The instantaneous channel gain being in a region, its representing symbol is sent back and the transmitter selects the codewords transmission parameters such that the system performance is optimized. The ARQ, on the other hand, is a well-known approach applied in today's networks to increase the transmission reliability in the absence of the transmitter CSI. From an information-theoretic point of view, the ARQ systems can be viewed as channels with sequential feedback where the transmitter CSI is refined in the retransmissions based on the message decoding status. In a general ARQ approach, the transmitter considers some initial transmission rate and power with no pre-knowledge about the channel quality. Then, with the help of ARQ, the decoding status at the receiver will be reported back to the transmitter via one bit feedback. Based on the received feedback, it is decided by the transmitter whether to retransmit the …