LAUSR

This paper presents a joint calculation of the precoding matrix indicator (PMI) and rank indicator (RI) in multiple-input multiple-output (MIMO) downlink third generation partnership project (3GPP) universal mobile telecommunications system (UMTS)/long-term evolution (LTE) systems. The standard codebook based precoding is used where the mobile station estimates and sends some measures related to the channel state information to the base station for link adaptation purpose. We study the system performance using link-level simulation under realistic assumptions. Specifically, we use K-best MIMO decoding in the presence of perfect and imperfect estimation of the channel and the noise variance, as well as white and colored noise plus interference defined by the 3GPP LTE specification documents. Similar to some previous results, our method for calculating the PMI/RI is based on maximizing the mutual information between the transmitted and received symbols with respect to the precoding matrix. We evaluate the system performance using two channel measures when computing the mutual information. The first measure is the estimated channel whereas the second one is the estimated channel covariance matrix. Although these two measures have been used in different literature, there have not been conclusive results on which one should be used in different practical scenarios. We assess the system performance using Monte Carlo simulations and compare different test cases, e.g., 3GPP channel models, wideband, and subband PMI/RI calculation, white/color noise plus interference, etc. We then conclude our work with technical recommendations for different scenarios.