MIMO-OFDM系統之指數有效訊雜比映射計算

Abstract

指數有效訊雜比(SNR)映射(EESM)技術已經是廣泛地被使用在OFDM系統中來決定實際應用的調變和編碼組合或系統模擬的錯誤率估測中。EESM是將一在多路徑通道下OFDM系統的錯誤率對應到加成性白雜訊高斯通道的錯誤率。對單輸入單輸出(SISO)OFDM系統而言，EESM應用的容易而且效能也令人滿意。然而，對多輸入多輸出(MIMO)OFDM系統而言，EESM的應用會變得沒那麼的直接，且沒有唯一的作法。在本論文中，我們研究適用於最大似然偵測器(ML)的MIMO-OFDM系統的EESM計算。眾所皆知，ML偵測器的錯誤率可由自由距離(free distance)來計算。然而，自由距離的計算需要很高的計算複雜度。我們因此提出使用一個基於QR分解的的方法來推導出一個下限值，使用這個下限值，我們發展出一個新的EESM計算方法。模擬結果顯示所提的方法不僅效能比傳統的EESM好而且擁有較低的計算複雜度。

Exponential effective signal-to-noise-ratio (SNR) mapping (EESM) is widely used technique for OFDM systems to determine a modulation and coding set in real-time operation or predict error rate in system simulations. EESM maps the error rate of an OFDM system with a multipath channel to that with an additive white noise channel. For single-input-single -output (SISO) OFDM systems, the conduction of EESM is straightforward and its performance is satisfactory. For multi-input-multi-output (MIMO) OFDM systems, however, EESM computation is much more involved and there is no unique way to do the job. In this thesis, we study EESM computation for the MIMO-OFDM system with the maximum likelihood (ML) detector. It is known that the error rate of the ML detector can be calculated by the free distance. However, the calculation of the free distance requires high computational complexity. We then propose using a QR-decomposition-based method to derive a lower bound. Using the lower bound, we develop a new EESM computation method. Simulations show that the proposed method not only outperforms the conventional method, but also has lower computational complexity.