High throughput wireless downlink packet data access with multiple antennas and multi user diversity.

摘要

Wireless communication channels with multiple antennas at both the transmitter and the receiver, or multiple-input multiple-output (MIMO) systems, have been recognized as one of the most prominent enablers of future generation telecommunications systems. Recent advances in multiuser communications that exploit multiuser diversity with scheduling algorithms, adaptive coding and modulation, and automatic repeat request algorithms, have proven the high efficiency of multiuser single-input single-output systems. This thesis contributes to the knowledge of the capacity of multiuser multiple antenna systems, and more specifically of the MIMO broadcast channel, with scheduling algorithms and rate adaptation. A novel analysis is provided to study the optimal number of users that should be allocated power in order to achieve the sum-capacity of MIMO broadcast channels, as well as the optimal power allocation and the optimal transmitter covariance matrices in the asymptotically high power region. Cases where receivers are equipped with a single or with multiple antennas are considered, and the fundamental differences between these systems are discussed. It is shown that intuition can sometimes be deceptive and extensive examples are provided to illustrate our findings. This analysis is then applied to N-user scheduling algorithms for throughput maximization, with the additional goal of providing low-complexity solutions. Similarities and differences with receive antenna selection algorithms are discussed. N-user scheduling algorithms are also studied in the context of sub-optimal transmitter-based linear spatial multiplexing schemes with complete channel state information at the transmitter. A novel interference-avoidance scheme is proposed with only partial channel state information available at the transmitter. Both throughput maximization and proportionally fair scheduling are considered. We provide analytical results when possible. Simulations are used to illustrate our analysis, and to study the performance of transmission schemes and scheduling algorithms when analysis is too complex. Eventually future directions for possible research are given.