MIMO Wireless Communications Under Statistical Queueing Constraints

The performance of multiple-input multiple-output (MIMO) wireless systems is investigated in the presence of statistical queueing constraints. Queuing constraints are imposed as limitations on buffer violation probabilities. The performance under such constraints is captured through the effective capacity formulation. A detailed analysis of the effective capacity is carried out in the low-power, wideband, and high signal-to-noise ratio (SNR) regimes. In the low-power analysis, expressions for the first and second derivatives of the effective capacity with respect to SNR at SNR = 0 are obtained under various assumptions on the degree of channel state information at the transmitter. Transmission strategies that are optimal in the sense of achieving the first and second derivatives are identified. It is shown that while the first derivative does not get affected by the presence of queueing constraints, the second derivative gets smaller as the constraints become more stringent. Through the energy efficiency analysis, this is shown to imply that the minimum bit energy requirements do not change with more strict limitations but the wideband slope diminishes. Similar results are obtained in the wideband regime if rich multipath fading is being experienced. On the other hand, sparse multipath fading with bounded number of degrees of freedom is shown to increase the minimum bit energy requirements in the presence of queueing constraints. Following the low-SNR study, the impact of buffer limitations on the high-SNR performance is quantified by analyzing the high-SNR slope and the power offset in Rayleigh fading channels. Finally, numerical results are provided to illustrate the theoretical findings, and to demonstrate the interactions between the queueing constraints and spatial dimensions over a wide range of SNR values.