多輸入多輸出正交分頻多工系統之子載波間干擾連續消除

Abstract

OFDM系統中通道快速變化會破壞子載波間之正交性，並因而導致子載波間干擾（intercarrier interference, ICI）。在單輸入單輸出（single-input single-output, SISO）-OFDM系統中，用於緩和ICI問題的方法有最小均方誤差（Minimum mean square error, MMSE）等化器與連續ICI消除器（successive ICI cancellation），而這兩種方法效能主要受限於通道估計的準確度，要獲得準確的通道估計值需要大量的pilot tones，但是相對付出代價是會減少頻寬使用效率。在本論文中，吾人針對MIMO-OFDM系統中ICI的問題提出一個遞迴估計通道與連續消除ICI的演算法，其基本概念為連續消除ICI並遞迴重新估計通道。MIMO-OFDM系統中會將所有的子載波一起排序供後續作ICI消除，此外，吾人在MMSE估計上會利用快速演算法以減少運算複雜度。最後模擬結果顯示出，本論文提出的演算法作用於MIMO-OFDM系統中通道快速變化的環境可以大幅降低ICI效應。

Fast channel variation will destroy the orthogonality of subcarriers in OFDM systems and this will result in intercarrier interference (ICI). Minimum mean square error (MMSE) and successive ICI cancellation have been proposed to mitigate this problem in single-input single-output (SISO) OFDM systems. However, the performance of these methods greatly depends on the channel estimation. To obtain good estimation, a great amount of pilot tones are required and this will reduce the bandwidth utilization efficiency. In this thesis, we propose a recursive channel estimation and successive ICI cancellation for MIMO-OFDM systems. The main idea is to successively cancel ICI and perform channel re-estimation. In ICI canceling, all tones in MIMO-OFDM systems are jointly sorted. To reduce the computational complexity, we also employ a fast algorithm in the MMSE estimate. Simulations show that the proposed algorithm can greatly reduce the ICI effect in MIMO-OFDM systems operated in the fast channel variation environment.