TY - GEN
T1 - Multiple-input multiple-output Gaussian channels
T2 - 2008 IEEE Information Theory Workshop, ITW
AU - Rodrigues, Miguel R.D.
AU - Pérez-Cruzy, Fernando
AU - Verdú, Sergio
PY - 2008
Y1 - 2008
N2 - We investigate the input covariance that maximizes the mutual information of deterministic multiple-input multipleoutput (MIMO) Gaussian channels with arbitrary (not necessarily Gaussian) input distributions, by capitalizing on the relationship between the gradient of the mutual information and the minimum mean-squared error (MMSE) matrix. We show that the optimal input covariance satisfies a simple fixedpoint equation involving key system quantities, including the MMSE matrix. We also specialize the form of the optimal input covariance to the asymptotic regimes of low and high snr. We demonstrate that in the low-snr regime the optimal covariance fully correlates the inputs to better combat noise. In contrast, in the high-snr regime the optimal covariance is diagonal with diagonal elements obeying the generalized mercury/waterfilling power allocation policy. Numerical results illustrate that covariance optimization may lead to significant gains with respect to conventional strategies based on channel diagonalization followed by mercury/waterfilling or waterfilling power allocation, particularly in the regimes of medium and high snr.
AB - We investigate the input covariance that maximizes the mutual information of deterministic multiple-input multipleoutput (MIMO) Gaussian channels with arbitrary (not necessarily Gaussian) input distributions, by capitalizing on the relationship between the gradient of the mutual information and the minimum mean-squared error (MMSE) matrix. We show that the optimal input covariance satisfies a simple fixedpoint equation involving key system quantities, including the MMSE matrix. We also specialize the form of the optimal input covariance to the asymptotic regimes of low and high snr. We demonstrate that in the low-snr regime the optimal covariance fully correlates the inputs to better combat noise. In contrast, in the high-snr regime the optimal covariance is diagonal with diagonal elements obeying the generalized mercury/waterfilling power allocation policy. Numerical results illustrate that covariance optimization may lead to significant gains with respect to conventional strategies based on channel diagonalization followed by mercury/waterfilling or waterfilling power allocation, particularly in the regimes of medium and high snr.
KW - Gaussian noise
KW - MMSE
KW - Multiple-input multiple-output systems
KW - Mutual information
KW - Optimal input covariance
UR - http://www.scopus.com/inward/record.url?scp=51849124625&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=51849124625&partnerID=8YFLogxK
U2 - 10.1109/ITW.2008.4578704
DO - 10.1109/ITW.2008.4578704
M3 - Conference contribution
AN - SCOPUS:51849124625
SN - 9781424422708
T3 - 2008 IEEE Information Theory Workshop, ITW
SP - 445
EP - 449
BT - 2008 IEEE Information Theory Workshop, ITW
Y2 - 5 May 2008 through 9 May 2008
ER -