Antenna subset selection in MIMO wireless systems

Multiple input multiple output (MIMO) technology improves the throughput and reliability in wireless systems. A key factor limiting the widespread deployment of this technology is the cost of the RF hardware associated with additional antennas. There is therefore a need for low cost low complexity techniques that deliver the leverages of MIMO technology. Antenna subset selection where transmission/reception is performed through a subset of the available antenna elements is an attractive solution.; In this thesis we study antenna selection (AS) for MIMO wireless systems. First we discuss AS for maximizing the channel capacity. We propose a suite of near-optimal AS algorithms, one of which in particular enables a tractable statistical analysis. We show that the capacity with AS is statistically lower bounded by the capacity of a set of parallel, independent SIMO channels with selection combining on each link. This result provides the crucial step in proving the equivalence of diversity order between the MIMO system with the full set of antennas and the MIMO system with AS. It also allows us to link up maximum ratio combining and selection in SIMO systems and Foschini's representation for capacity (full channel) and the capacity with AS in MIMO systems.; The AS algorithms and analysis depend to a large extent on the space-time signaling scheme employed on the link. AS based on maximizing capacity may prove detrimental to performance when suboptimal signaling/receiver schemes are employed over the link. This implies the need for tailoring the selection to the requirements of the signaling and receiver schemes. We develop AS algorithms and performance analysis for orthogonal space-time block coding and spatial multiplexing with MMSE and OSIC receivers.; Finally, we discuss AS when only channel statistics are known. This type of selection offers array gain but does not deliver the diversity benefits possible with exact channel knowledge. We present performance analysis for space-time coding and spatial-multiplexing schemes in correlated channels and use these results to develop AS algorithms for known channel statistics.