Performance of MIMO E-SDM Systems Using Channel Prediction in Actual Time-Varying Indoor Fading Environments

Huu Phu BUI  Hiroshi NISHIMOTO  Toshihiko NISHIMURA  Takeo OHGANE  Yasutaka OGAWA  

IEICE TRANSACTIONS on Communications   Vol.E91-B   No.6   pp.1713-1723
Publication Date: 2008/06/01
Online ISSN: 1745-1345
DOI: 10.1093/ietcom/e91-b.6.1713
Print ISSN: 0916-8516
Type of Manuscript: Special Section PAPER (Special Section on 2007 International Symposium on Antennas and Propagation)
Category: Smart Antennas & MIMO
MIMO systems,  eigenbeam-space division multiplexing (E-SDM),  transmit beamforming,  channel prediction,  time-varying fading environments,  

Full Text: PDF>>
Buy this Article

In time-varying fading environments, the performance of multiple-input multiple-output (MIMO) systems applying an eigenbeam-space division multiplexing (E-SDM) technique may be degraded due to a channel change during the time interval between the transmit weight matrix determination and the actual data transmission. To compensate for the channel change, we have proposed some channel prediction methods. Simulation results based on computer-generated channel data showed that better performance can be obtained when using the prediction methods in Rayleigh fading environments assuming the Jakes model with rich scatterers. However, actual MIMO systems may be used in line-of-sight (LOS) environments, and even in a non-LOS case, scatterers may not be uniformly distributed around a receiver and/or a transmitter. In addition, mutual coupling between antennas at both the transmitter and the receiver cannot be ignored as it affects the system performance in actual implementation. We conducted MIMO channel measurement campaigns at a 5.2 GHz frequency band to evaluate the channel prediction techniques. In this paper, we present the experiment and simulation results using the measured channel data. The results show that robust bit-error rate performance is obtained when using the channel prediction methods and that the methods can be used in both Rayleigh and Rician fading environments, and do not need to know the maximum Doppler frequency.