Universal approach to performance optimization of multiuser MIMO systems

Multiple antennas have been applied in order to increase the performance and reliability of single- and multiuser wireless systems. In this work, we study the behavior of a large class of performance metrics and their corresponding optimal transmission strategies under individual and sum power constraints for perfect CSI at the base and mobiles. At first, we derive the unique optimal single-user transmit strategy which is generalized water-filling. As a result, we show that at low SNR values only one user who has highest largest channel eigenvalue is active. The performance of the up- and downlink multiuser system can be measured by different performance metrics, e.g. the sum capacity if optimum decoding and successive interference cancellation (SIC) is applied. Another performance metric is the average sum MSE of the uplink channel if the linear multiuser MMSE receiver is applied. Both performance metrics can be written as the trace of a matrix-monotone function.