Signal Processing in Arrayed MIMO Systems

Multiple-Input Multiple-Output (MIMO) systems, using antenna arrays at both receiver and transmitter, have shown great potential to provide high bandwidth utilization efficiency. Unlike other reported research on MIMO systems which often assumes independent antennas, in this thesis an arrayed MIMO system framework is proposed, which provides a richer description of the channel charac- teristics and additional degrees of freedom in designing communication systems. Firstly, the spatial correlated MIMO system is studied as an array-to-array system with each array (Tx or Rx) having predefined constrained aperture. The MIMO system is completely characterized by its transmit and receive array man- ifolds and a new spatial correlation model other than Kronecker-based model is proposed. As this model is based on array manifolds, it enables the study of the effect of array geometry on the capacity of correlated MIMO channels. Secondly, to generalize the proposed arrayed MIMO model to a frequency selective fading scenario, the framework of uplink MIMO DS-CDMA (Direct- Sequence Code Division Multiple Access) systems is developed. DOD estimation is developed based on transmit beamrotation. A subspace-based joint DOA/TOA estimation scheme as well as various spatial temporal reception algorithms is also proposed. Finally, the downlink MIMO-CDMA systems in multiple-access multipath fading channels are investigated. Linear precoder and decoder optimization problems are studied under different criterions. Optimization approaches with different power allocation schemes are investigated. Sub-optimization approaches with close-form solution and thus less computation complexity are also proposed