Optimization algorithms for transmit power minimization based on partial channel state information in mimo systems

Abstract

Although frequency spectrum is infinite, only a very limited range of spectrum can be utilized for wireless communications with current technology. As wireless communications evolve to the Fourth Generation (4G), frequency utilization becomes a challenge since much higher data rate is expected. Frequency spectrum for new wireless systems is scarce and expensive. It is desired to maximize the transmission rate within limited bandwidth. This demand can not be met without a significant change in spectral efficiency. Transmission over Multiple-Input Multiple-Output (MIMO) systems most promising solution. Multiple antennas are employed at both transmitter and receiver, using the same frequency spectrum. In this thesis, Orthogonal Frequency Division Multiplexing (OFDM) systems in combination with MIMO transmission are considered. Only partial channel state information (CSI) at transmitter is assumed. The transmitter design with coder-beamformer is adopted to increase system robustness against uncertainties of statistical CSI. Since optimal bit allocation is of high complexity, this thesis proposes a novel algorithm for minimizing transmit power in MIMO-OFDM systems. The proposed scheme has better performance and lower complexity than existing algorithm when average number of bits per sub-channel is small with respect to constellation size.