正交分頻多工系統中的IQ不平衡效應與載波頻率偏移的聯合偵測與補償之研究

Abstract

在本論文中，提出了一個結合載波頻率偏移與IQ不平衡效應的基頻IQ偵測與補償演算法。在目前已知的方法中，通常存在有下列兩種缺點：收斂速度慢以及補償範圍無法符合規格的要求。因此我們提出了一個可以偵測與補償增益錯誤1 dB，相角錯誤10度與載波頻率偏移50 ppm在2.4 GHz的頻帶上的架構。在這樣的條件下，整體系統的效能存在有少於0.8dB封包錯誤率的設計損失並且比參考的演算法有高達6dB的系統效能提升。此外，一個基於虛擬載波頻率偏移技巧的載波頻率偏移演算法被提出來。此演算法適用於存在有IQ不平衡效應的情況下並且可以容忍2dB的增益錯誤與20度的相角錯誤在多重路徑的環境中。此演算法只利用三個連續的訓練序列來萃取載波頻率偏移的大小。從模擬的結果中，演算法的偵測錯誤量大約是0.3 ppm並且小於傳統載波頻率偏移偵測演算法，因此所提出的演算法有機會實現高效能的接收端。另外，我們所提出的虛擬載波頻率偏移演算法除了可以相容於傳統的演算法之外並且適合用在硬體的實現上。

In this thesis, a baseband I/Q estimation and compensation algorithm is proposed which considers the joint effects of carrier frequency offset (CFO) and I/Q mismatch (IQ-M). Current solutions converge too slowly in the estimation, or the compensation range of gain-phase error and CFO cannot meet the required values. The proposed algorithm uses a feed-forward architecture which is able to estimate and compensate the non-ideal effect up to gain error 1 dB, phase error 10 degree, and CFO 50 ppm at carrier frequency 2.4 GHz. Under this condition, the system performance has less than 0.8 dB design loss in terms of packet error rate, and it is shown to have more than 6 dB improvement compared with the reference designs. By the way, a novel CFO estimation algorithm based on pseudo CFO (P-CFO) is developed to estimate the CFO value under the conditions of IQ-M for direct conversion structures with 2 dB gain error and 20-degree phase error in frequency selective fading channels. This algorithm only uses three consecutive training symbols to extract the CFO value with the technique of signal processing. From simulation results, it is shown that the estimation error of the proposed method is about 0.3 ppm which is smaller than the conventional method based on two-repeat preamble to achieve high performance receivers. In addition, the proposed P-CFO algorithm is also compatible with the conventional method and suitable for implementation issues.