使用基底擴展模型實現OFDM通道估測之研究

Abstract

在無線通訊傳輸環境下，為了降低位元錯誤率，通道估測是一項重要的議題。在本篇論文中，我們採用了近年受到廣泛地使用的正交分頻多工(OFDM)系統，利用傑克模型來模擬雙重選擇通道環境，藉由基底擴展模型來實現通道估測的任務。我們將分析常見的基底擴展模型，並推導出理論的估測錯誤率，以及使用維納濾波器來提高已估測通道之訊雜比，可運用在不同的基底擴展模型上。同時也考慮了實際環境中，載波頻率偏移以及相位雜訊的影響，我們提出改善的模型藉此降低其效應。另外，為了降低運算複雜度以及提高收斂速率，我們提出了具有不同調整步伐的最小均方演算法(LMS)以及遞迴最小平方演算法(RLS)，理論分析出最佳的非時變調整步伐，在不同的投影係數上，其調整步伐將會與其基底的重要性呈現正比關係。此外。如果當都普勒頻率在低頻區間改變的情況下，基底也將隨之改變，我們則提出藉由少量的參數記憶以及冪法演算法來模擬追蹤出改變後的基底，可有效節省運算複雜度或記憶空間的使用。

For reducing the bit-error rate (BER) in wireless communication systems, it is important to provide a channel estimator at receiver in orthogonal frequency-division multiplexing (OFDM) system. We use the basis-expansion model (BEM) to estimate the doubly-selective channel which is simulated by Jake's model. We would analyze and discuss some often seen BEMs, thus, we could derive the theoretical MSE and apply the Wiener filter to enhance the signal-to-noise ratio (SNR) for estimated channel. The carrier frequency offset (CFO) and phase noise (PHN) issues are also considered in our proposed scheme. Moreover, for reduce complexity, we update the BEM coefficients by different weighting corresponding to the significance of basis, which could be verify by the theoretic derivation. Besides, when Doppler frequency changes, we propose two methods to obtain the new bases, these two schemes use Power method and asymptotic sinusoidal function that need only few parameters and computation than conventional methods.