Logo image
多目標強健波束成型設計之目標SINR追蹤與QoS於不完全通道資訊之多輸入多輸出無線系統
Thesis

多目標強健波束成型設計之目標SINR追蹤與QoS於不完全通道資訊之多輸入多輸出無線系統

Chen, Wei Yu
Masters, 國立清華大學, 電機工程學系
2016

Abstract

線性矩陣不等式 多細胞多輸入多輸出 半正定規劃 強健H-infinity設計 服務質量 能量消耗 多目標最佳化問題 Linear matrix inequality (LMI) Multicell multi-input multi-output (MIMO) Semidefinite programming (SDP) Robust H-infinity design scheme Quality of service (QoS) Power consumption Multiobjective optimization problems (MOPs)
Although high Quality of Service (QoS) requirement of the multicell multi-input multi-output (MIMO) interference wireless network can be achieved by joint processing (JP) and coordinated beamforming (CB), it is difficult to deal with the delay effect of the system and the uncertainty of the channel information. In this study, we propose a multiobjective beamforming scheme with optimal signal-to-interference-plus-noise ratio (SINR) tracking methods to achieve QoS with minimum power consumption for multicell MIMO interference wireless system with imperfect channel coefficients. The advantages of the proposed multiobjective beamforming scheme are threefold. First, the designed beamforming can achieve three objectives as follows: optimal H-infinity SINR tracking performance, minimum total power consumption, and minimum worst case interference in the multicell MIMO system simultaneously. Second, we show that the multicell interference, round trip delay, channel fading and noise can be efficiently weakened by the proposed robust H-infinity SINR tracking scheme to provide a better feasibility than the conventional methods. Third, through the cooperation between SINR tracking model and the weighting sum semidefinite programming (SDP) with multiobjective evolution algorithm (MOEA), the Pareto optimal beamforming weights can be obtained to achieve QoS requirements with minimum power consumption simultaneously. Finally, we provide several numerical examples to show that the proposed design scheme can achieve favorable results and guarantee the theoretical analysis.

Metrics

1 Record Views

Details

Logo image