On the Use of Jensen’s Inequality for Mimo Channel Capacity Estimation

In this paper, we derive a universal upper bound on the MIMO channel capacity, which is not limited to a particular scenario, using the correlation matrix approach and the Jensen’s inequality. This bound accounts for both transmit and receive branch correlation in such a way that the impact of these branches can be estimated separately, which simplifies the procedure substantially. Some simple analytical results, which quantify the impact of correlation on the MIMO capacity in an explicit way, are given. We show that correlation increase is equivalent to SNR decrease in some cases. Finally, we discuss the paradox of zero correlation and provide a statistical explanation for it. We demonstrate why zero mean correlation is not a guarantee of high capacity. Mean magnitude or RMS correlation should be used for the capacity estimation of degenerate MIMO channels.