TY - GEN
T1 - Angular domain processing for MIMO wireless systems with non-uniform antenna arrays
AU - Huang, Dayu
AU - Raghavan, Vasanthan
AU - Poon, Ada S.Y.
AU - Veeravalli, Venugopal V.
PY - 2008
Y1 - 2008
N2 - Many recent works have proposed spatial channel models for MIMO applications in the 1 to 10 GHz band. However, recent attention has shifted to the 60 GHz regime where the opening up of the spectrum has created new opportunities for multi-antenna communications. With a broad focus on the 60 GHz regime, we revisit MIMO channel modeling in the classical setting and propose an angular domain decomposition of the channel in this work. In the special case of non-uniform linear arrays, this proposal corresponds to non-uniformly partitioning the angular domain commensurate with the non-uniformity structure of the arrays and estimating the channel gains in each partition with a multi-window spectral estimator. We show that the optimal choice of basis functions that minimize the worst-case bias of the spectral estimator are solutions to certain generalized eigenvalue problems. Based on this structure, we develop a spatial model that generalizes existing MIMO channel modeling paradigms.
AB - Many recent works have proposed spatial channel models for MIMO applications in the 1 to 10 GHz band. However, recent attention has shifted to the 60 GHz regime where the opening up of the spectrum has created new opportunities for multi-antenna communications. With a broad focus on the 60 GHz regime, we revisit MIMO channel modeling in the classical setting and propose an angular domain decomposition of the channel in this work. In the special case of non-uniform linear arrays, this proposal corresponds to non-uniformly partitioning the angular domain commensurate with the non-uniformity structure of the arrays and estimating the channel gains in each partition with a multi-window spectral estimator. We show that the optimal choice of basis functions that minimize the worst-case bias of the spectral estimator are solutions to certain generalized eigenvalue problems. Based on this structure, we develop a spatial model that generalizes existing MIMO channel modeling paradigms.
UR - http://www.scopus.com/inward/record.url?scp=70349658512&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=70349658512&partnerID=8YFLogxK
U2 - 10.1109/ACSSC.2008.5074791
DO - 10.1109/ACSSC.2008.5074791
M3 - Conference contribution
AN - SCOPUS:70349658512
SN - 9781424429417
T3 - Conference Record - Asilomar Conference on Signals, Systems and Computers
SP - 2043
EP - 2047
BT - 2008 42nd Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2008
T2 - 2008 42nd Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2008
Y2 - 26 October 2008 through 29 October 2008
ER -