以鄰近量化改善採有限回饋之多輸入多輸出廣播通信品質

Abstract

在多用戶多輸入多輸出下行系統中,基地台利用通道狀態資訊預編碼信號,使用戶 間干擾可以被消除。利用有限回饋的方式,可以實際使用通道狀態資訊。但是量化 誤差會產生用戶間干擾,進而造成傳輸通量的上限效應。在此論文中,我們提出一 個低複雜度的有限回饋方法,可以提供高解析度的通道狀態資訊而不須建立大碼簿; 這個方法利用鄰近通道矩陣製造出精練的通道狀態資訊。經由模擬結果和效能分析, 我們證實提出的方法得以較小複雜度改善通量上限效應。

In multiuser multiple-input multiple-output downlink (MIMO) systems, base stations precode signals in order to suppress the inter-user interference based on channel state information (CSI). Limited feedback enables the practical use of CSI to achieve this goal. The challenge with limited feedback is that the quantization errors create inter-user interference, which can cause through- put ceilings at high SNR. We propose a low-complexity limited feedback method to provide high resolution instead of constructing a large codebook. The method exploits adjacent channel matrix to create refined CSI. We also utilize it to improve scheduling and support multiple resolution. Through simulation results and performance analysis, we demonstrate that proposed method can improve throughput ceiling with lower complexity.