Exploiting Only Channel Spatial Correlations for Optimal Power-Saving Input Covariance in MIMO-Based Wireless Systems

Jung-Chieh CHEN  Chao-Kai WEN  

IEICE TRANSACTIONS on Communications   Vol.E90-B   No.12   pp.3505-3513
Publication Date: 2007/12/01
Online ISSN: 1745-1345
DOI: 10.1093/ietcom/e90-b.12.3505
Print ISSN: 0916-8516
Type of Manuscript: PAPER
Category: Fundamental Theories for Communications
capacity,  channel state information,  convex optimization,  large system,  MIMO,  power minimization,  spatial correlation,  

Full Text: PDF(290.1KB)>>
Buy this Article

Most studies into multiple-input multiple-output (MIMO) antenna systems have aimed at determining the capacity-achieving (CA) input covariance given a certain degree of channel state information (CSI) at the transmitter and/or the receiver side. From the practical perspective, however, there is a growing interest in investigating the scenario where the system performance is power-limited as opposed to rate-limited. Of particular concern is the open problem of solving the optimal power-saving (PS) input covariance for spatially correlated MIMO channels when only the long-term (slow-varying) channel spatial covariance information is available at the transmitter. In an attempt to achieve this goal, this paper analyzes the characteristics of the optimal PS input covariance given the knowledge of channel spatial covariance information and the rate constraint of the transmission. Sufficient and necessary conditions of the optimal PS input covariance are derived. By considering the large-system regimes, we further devise an efficient iterative algorithm to compute the asymptotic optimal PS input covariance. Numerical results will show that the asymptotic solution is very effective in that it gives promising results even for MIMO systems with only a few antennas at the transmitter and the receiver.