Performance of the MIMO downlink channel with multi-mode adaptation and scheduling

This paper presents an algorithm for switching transmission methods (whether for diversity or data rate maximization) over a multiple antenna broadcast channel. The proposed approach exploits long-term spatial selectivity of the user channels to decide between different diversity and multiplexing modes. We investigate the performance of this adaptive transmission method with different scheduling polices (both sensitive to user queue sizes and independent of user queue sizes). While the methods outlined in this paper are general, we present our results in the context of a broadcast channel with four transmit antennas at the central transmitter, and four antennas at each receiver. Our results indicate a gain of approximately 20% in average system throughput with the proposed algorithm over realistic channel models.