雙選擇性通道下正交分頻多工系統中使用馬可夫鏈蒙地卡羅法之基於期望值最大化接收機設計

Abstract

雙選擇性(衰減)通道造成載波間干擾(Inter-carrier Interference, ICI)的問題並降低系統效能，受到鄰近子載波之能量擴散而產生的載波間干擾效應，是本論文首要的估測目標，因而提出一種正交分頻多工(Orthogonal Frequency Division Multiplexing, OFDM)系統接收機，其為基於期望值最大化演算法(Expectation-Maximization, EM)設計而成。在頻域模式下進行系統分析，將EM演算法與馬可夫鏈蒙地卡羅法(Markov chain Monte Carlo, MCMC)結合，並以最大相似度(Maximum Likelihood, ML)作為判定準則，我們得到一套有系統地估測載波間干擾的方法，稱之為「EM通道估測法」。此外，為了減低運算複雜度，以及充分利用時變通道所賦予的時間多樣性，ML-EM接收機採用「分群式載波間干擾消除器」，針對通過此消除器之後的接收信號，估算適當的殘餘載波間干擾功率，以提高資料偵測的正確率。電腦模擬結果顯示，相較於以往一階等化器，我們提出的ML-EM接收機在錯誤率方面的表現，有著明顯的進步。

Doubly selective (fading) channels cause the inter-carrier interference (ICI) problem and thus degrade the system performance. In order to estimate the ICI effect made by the spreading energy of adjacent subcarriers, we propose an expectation-maximization (EM)-based receiver for orthogonal frequency division multiplexing (OFDM) systems. In this paper, we use the frequency domain model for system analysis and derive the EM channel estimation method by combining the Markov Chain Monte Carlo (MCMC) method with the EM algorithm according to the maximum-likelihood (ML) criterion. Besides, the proposed EM-based receiver is incorporated with the group-wise ICI cancellation method for reducing computational complexity and exploiting the inherent time diversity in time-variant channels. After the ICI cancellation, the residual ICI power is calculated for the ICI-reduced signals with the goal of making data detection more correctly. Results of computer simulation demonstrate that the ML-EM receiver performs much better than the conventional one-tap equalizer.