On the impact of interference on data protocol performance in multicellular wireless packet networks with MIMO links

We study here the performance of the data transmission protocol in a multicell SDMA/TDMA MIMO system running a distributed throughput-driven slot allocation algorithm, as a function of reuse distance, propagation exponent, and network load. Multiple antenna elements are used at both transmitter and receiver side, and perfect channel state information (CSI) is available at the receiver only. A simplified scenario with two access points is considered. We find that the capacity of the network is heavily affected by the presence of interference and we show that the joint use of an estimation of interference covariance matrix and adaptive stream-limited antenna selection prevents capacity degradation when the offered load increases. The results show that the adaptive stream-limited antenna selection introduces a considerable performance improvement over the traditional approach of frequency reuse.