具關連性多輸入多輸出系統之位元分配與傳送器設計

Abstract

在本篇論文中，我們提出對於輸入端具關連性多輸入多輸出系統，其 位元配置與傳送器的設計方法。首先我們根據具關連性通道之錯誤變 異係數的統計特性推導出最佳位元配置與平均錯誤率下限，接著我們 設計出能最小化平均錯誤率下限之傳送器與最適合的通道數目。模擬 結果顯示我們所提出的方法可以有效降低具關連性多輸入多輸出系 統之錯誤率。

In this thesis we consider the design of statistical precoder and statistical bit allocation for multi-input multi-output (MIMO) systems over correlated channels. We assume the correlated channel is slow fading and full channel state information is available at the receiver, while only the statistics of the correlated channels is assumed to be known at transmitter. We will first derive the statistical bound of bit error rate (BER) and the corresponding optimal real bit allocation. Based on this statistical BER bound, the optimal unitary statistical precoder is derived both for linear and decision feedback receiver. Second, the statistical integer bit allocation is designed the greedy algorithm. Finally, different number of substreams will be considered and selected by statistical BER bound. Simulations show lower BER can be achieved when optimal number of substreams is selected for correlated channels.