to appear in Proc. of AP-S Intern. Symp., Monterey, California, June 20-26, 2004.
The capacity and error rate performance of MIMO (multiple-input multiple-output) wireless communication systems depends explicitly on the spatial correlation of the channel perceived by the physical layer . Characterization of the spatial correlations - essential for accurately predicting system performance - is challenging due to the interdependence between the propagation environment and the antenna geometry. One generic MIMO channel model that has emerged in a variety of applications of MIMO communication is the clustered channel model. In this model, scatterers are grouped in clusters, experiencing delay and angular spreads. Previous work has focused on evaluating performance in this channel under a uniform linear and circular array assumption. Other array geometries, despite potential implementation advantages, have been neglected. In this paper we study the impact of different array configurations in indoor propagation environments. We measure the capacity/diversity gain attainable by different array geometries with seven elements and fixed interelement spacing. To make the discussion more concrete, we base our simulations on the IEEE 802.11n Technical Group (TG) clustered channel model. Our results show that averaged with respect to cluster location, uniform linear arrays often yield the highest capacity/diversity gains. However, in poor scatterer environments and for compact arrays the „Starš configuration provides the best system performance.
This paper is available on .IEEE Xplore .