A reduced-rank Eigenbasis MIMO channel model

In designing optimal signaling strategies, the need for accurate yet simple channel models is ever-present. In this paper, we develop and investigate a reduced form of an analytical channel model using measured MIMO channel matrices. We investigate the use of this reduced form on predicting MIMO system performance for a square M-QAM closed loop MIMO system with a BER constraint. We compare the performance of the reduced model against both the full analytical channel model developed by Weichselberger and another popular analytical model known as the Kronecker model. Several different environments are explored to generate a range of coupling matrix performances. It is found that at low to moderate SNRs, the channel may be represented accurately by the new reduced model.