正交分頻多工存取系統上傳連結之載波頻率偏移估計與補償

Abstract

由於正交分頻多工存取技術（OFDMA）能夠在提供可靠多工存取的同時，也提供極佳的頻譜效益，以及對抗多路徑傳輸之影響，因此近年來已經吸引了廣泛的注目。不同於OFDM系統，OFDMA基地台會遭遇到因不同用戶端相異的載波頻率偏移（CFOs），所產生之嚴重的子載波之間干擾（ICI）與多用戶存取干擾（MAI）問題。載波頻率偏移估計之準確性會受到子載波之間干擾與多用戶存取干擾的影響。研究如何在OFDMA uplink系統上估計載波頻率偏移，以及消除CFO造成之干擾的演算法，將是本論文的重點 在本論文中，作者將提出兩種結合data-aided 載波頻率偏移估計演算法[3]與self-ICI cancellation [6]、received windowing [7]等干擾壓抑技術的CFO估測方法，使其能適用於OFDMA上傳連結系統。在估測完載波頻率偏移後，作者也提出兩種用於OFDMA上傳連結接收端之區塊載波頻率偏移補償演算法，分別為最小均方誤差區塊平行干擾消除（Minimum Mean Square Error Block Parallel Interference Cancellation，MMSE-BPIC），與區塊連續干擾消除（Block Successive Interference Cancellation，BSIC）演算法。經由模擬可得知，作者提出的載波頻率偏移估計演算法皆能提供接近[3]在單一用戶OFDM系統上所能達到的良好準確度；模擬結果也顯示，載波頻率偏移估計誤差對於我們提出的MMSE-BPIC與BSIC頻率補償演算法效能不會造成太大的影響，而且與其他頻率補償演算法相較，MMSE-BPIC與BSIC都能提供良好的效能表現。

In the recent years, the technique of orthogonal frequency division multiple access (OFDMA) has gained much attention because of the excellent bandwidth efficiency and the ability of combating multi-path effects while providing reliable multiple access capability. Different from the OFDM system, the OFDMA basestation has more severe inter-carrier interference (ICI) and the multiple access interference (MAI) problems which are caused by the multiple carrier frequency offsets (CFO) from different subscribers problem. The accuracy of the CFO estimation is critical to the impairment caused by ICI and MAI. The main subject of this thesis is the study of the CFO estimation and the CFO-introduced interference cancellation algorithms. In this thesis, we present two improved CFOs estimators which combine the data-aided CFO estimator [3] with two interference suppression techniques: the self-ICI cancellation [6] and the received windowing [7] methods. After having estimated the CFOs, we also propose two block-based algorithms for CFOs compensation at the uplink receiver of OFDMA systems. They are Minimum Mean Square Error Block Parallel Interference Cancellation (MMSE-BPIC) and Block Successive Interference Cancellation (BSIC). In the computer simulation, the proposed CFOs estimators can provide good accuracy which closes to that of the single-user OFDM system [3]. Simulation results show that the proposed MMSE-BPIC and BSIC algorithms are robust to the CFO estimation errors and can provide good performance compared with the other existing approaches.