Impact of multipath clustering on the performance of MIMO systems

Fading correlation has a profound impact on the MlMO system performance. Salz-Winters model is a popular tool to study this effect. However, it is limited to one cluster only. Measurements indicate that multipath often arrives in several clusters. We extend the Salz-Winters model to the case of multi-cluster channels and study it in detail. Closed-form expression for correlations are derived and applied to MIMO capacity/diversity gain analysis. The maximum gain/capacity are achieved provided that a minimum element spacing (derived in the paper) is respected. It is shown that the correlation has an oscillatory behavior as antenna spacing increases; the envelope of correlation is determined by a single cluster angular spread while the oscillations within the envelope are determined by the inter-cluster angular spread. It is demonstrated that the correlation depends significantly on the power distribution among the clusters. In the case of 2 widely-separated clusters, it is possible to orient the antenna array in such a way that the correlation is minimized and, hence, the capacity/gain are maximized. We study this optimization problem and derive the optimum array orientation. Overall, the paper presents a new insight on correlation properties of multipath clustered channels, and on the MIMO system performance over such channels.