Accurate Estimation of Correlation and Capacity for Hybrid Spatial-Angular MIMO Systems

Abstract

In this paper, several novel functions for accurately estimating the correlation and the multiple-input–multiple-output (MIMO) capacity of combined spatial and true polarization diversity (TPD) schemes are proposed for the first time. Minimum error estimation of the correlation for a hybrid spatial–TPD linear scheme previously reported cannot be obtained by simple nonlinear least-square estimators, and the use of genetic algorithms (GAs) provides an accurate solution. Channel matrix coefficients and MIMO capacities are analyzed for the diverse optimization strategies employed, which aim to identify the best prediction of MIMO performance. The novel functions are used with previously reported results and validated for Rayleigh fading scenarios with isotropic scattering using a multimode-stirred chamber (MIMO Analyzer). An accurate prediction of high correlation values is concluded to be of extreme importance for the final MIMO performance estimation. This has been found to be particularly important for estimating MIMO capacity with achieved prediction accuracies of 1.1% at SNR = 15 dB. In contrast, the accurate prediction of low correlation values has been found to have a less-important effect on the final capacity-predicting performance.