A Novel Low-Complexity Channel Estimation Approach for Single Carrier MIMO Frequency Selective Channels

Suyue LI  Jian XIONG  Lin GUI  Youyun XU  Baoyu ZHENG  

IEICE TRANSACTIONS on Communications   Vol.E96-B   No.1   pp.233-241
Publication Date: 2013/01/01
Online ISSN: 1745-1345
DOI: 10.1587/transcom.E96.B.233
Print ISSN: 0916-8516
Type of Manuscript: PAPER
Category: Wireless Communication Technologies
correlation channel estimation,  joint ISI and CCI cancellation,  MIMO training sequences,  LST-MMSE frequency domain equalization,  

Full Text: PDF(1.4MB)>>
Buy this Article

A simple yet effective time domain correlation channel estimation method is proposed for multiple-input multiple-output (MIMO) systems over dispersive channels. It is known that the inherent co-channel interference (CCI) and inter-symbol interference (ISI) coexist when the signals propagate through MIMO frequency selective channels, which renders the MIMO channel estimation intractable. By elaborately devising the quasi-orthogonal training sequences between multiple antennas which have constant autocorrelation property with different cyclic shifts in the time domain, the interferences induced by ISI and CCI can be simultaneously maintained at a constant and identical value under quasi-static channels. As a consequence, it is advisable to implement the joint ISI and CCI cancelation by solving the constructed linear equation on the basis of the correlation output with optional correlation window. Finally, a general and simplified closed-form expression of the estimated channel impulse response can be acquired without matrix inversion. Additionally, the layered space-time (LST) minimum mean square error (MMSE) (LST-MMSE) frequency domain equalization is briefly described. We also provide some meaningful discussions on the beginning index of the variable correlation window and on the cyclic shift number of m-sequence of other antennas relative to the first antenna. Simulation results demonstrate that the proposed channel estimation approach apparently outperforms the existing schemes with a remarkable reduction in computational complexity.