Antenna selection for MIMO systems based on information theoretic considerations

The intrinsic capacity formulation generates optimal transmit current and receive field sampling distributions that typically are impractical to realize physically and require complex transmit/receive hardware. A more practical scenario would be to deploy large, reconfigurable arrays and select an optimal or near-optimal subset of the antennas for connection to the (fewer available) transmit and receive hardware chains. This paper presents algorithms based upon mutual information quantities derived from the intrinsic capacity computation that can efficiently and effectively identify good choices of antennas. While the algorithms do not guarantee optimal antenna selection, results obtained using realistic channel models reveal the excellent performance of the techniques.