Frequency domain processing for multiple input multiple output channels

In the past two decades wireless communication technology has developed tremendously. Therefore, the techniques that can support high data rates and provide high spectral efficiencies are of great interest. Employing multiple antennas at both receiver and transmitter and space-time processing techniques has been shown to be able to provide high spectral efficiencies. The fundamental idea is to use multiple antennas combined with signal processing techniques at the receiver in order to increase the spectral efficiencies. The focus of this thesis is on frequency domain transmitter and receiver design for severe frequency selective channels which can effectively mitigate the effect of ISI and interference for single carrier wireless systems with multiple transmit and receive antennas.First, two new frequency-domain receivers for severe frequency-selective MIMO channels are introduced. These receivers have a frequency domain feedforward and a time domain feedback filter. Assuming the availability of perfect knowledge of the channel at the receiver, the performance of the proposed receivers are evaluated for different channel models. However, these receivers are non-linear and suffer from error propagation. The comparison of the performance of the proposed receivers with OFDM is also given.Second, a new iterative receiver is proposed for severe time-dispersive MIMO channels. This receiver has a frequency domain feedforward and feedback filter. Since the feedback loop takes into account not just the hard-decisions for each block but also the overall block reliability, the error propagation problem is significantly reduced.The problem of joint transmitter and receiver design for MIMO channels under the MSE criterion and with the perfect knowledge of the channel at the transmitter when the system does not have a priori bandwidth constraint is considered. It is proven that the optimum space-time filters are strictly bandlimited to a frequency set. The aggregated bandwidth of this set is at most the number of the substreams times the symbol rate. Next, we propose a linear frequency domain pre-equalization scheme for spatially correlated MIMO channels. In this scheme partial knowledge of the channel is available at the transmitter. To reduce the complexity the proposed precoder is implemented in the frequency domain. Finally, we introduce a power optimization algorithm for V-BLAST system with two transmit antennas. This algorithm minimizes the exact probability of vector error.