A Simple Multiuser MIMO-SDMA System in the Downlink Using Statistical Precoding
Jue Wang, Shi Jin, Xiqi Gao, Kai‐Kit Wong · UCL Discovery (University College London) · 2011
This paper proposes a new space-division multipleaccess (SDMA) approach for multiuser multiple-input multipleoutput downlink systems using statistical precoding. The proposed scheme makes use of rotated discrete Fourier transform (DFT) matrices to realize spatial divisions, each of which defines a cluster within which users have their transmit-side eigenmodes aligning closely with a common rotated DFT matrix. Users inside a cluster are divided into groups, and users exhibiting orthogonal eigen-directions are put in the same group realizing SDMA. It is shown that the proposed scheme can greatly enhance the average system throughput if compared with time-division multiple-access, especially when transmit-side correlation is high. Meanwhile, the proposed scheme outperformances traditional linear precoding schemes such as zero forcing when the feedback bits for the channel state information are limited. © VDE VERLAG GMBH.