Improved Channel Estimation in OFDM Wireless Communication Systems

In this dissertation, two efficient channel estimation schemes in orthogonal frequency division multiplexing (OFDM) wireless communication systems are proposed to improve the performance of conventional estimation techniques. Firstly, we propose an improved low-complexity channel estimation scheme in OFDM systems with transmit diversity by effectively eliminating inter-antenna interference (IAI). For decision-directed channel estimation (DDCE), the high complexity is caused by the joint estimation of the channels between the receiver and all transmitters. After successfully removing IAI, each channel can be estimated independently, which results in a considerable reduction of complexity. The existing decoupling process is performed by exploiting the correlation of channel frequency response (CFR) between adjacent subcarriers; however, partial IAI is remained due to the channel frequency selectivity. We propose a new algorithm to improve channel estimation performance which includes a new decoupling technique with weaker assumptions about the correlation of CFRs and residual IAI cancellation by using the results of DDCE and the latest estimates. In the existing decoupling process, noise components in the decoupled signal are correlated with transmitted data. When non-constant envelope modulation is employed, the standard least-square (LS) criterion for channel estimation is suboptimal due to the unequal noise variance at each subcarrier. To improve estimation accuracy, we adopt the best linear unbiased estimation (BLUE) method to de-emphasize the decoupled data with large noise variance. After combing with the proposed decoupling and IAI compensation, simulation results illustrate that the performance of the proposed algorithm is very close to the high complexity algorithm which jointly estimates all channel responses. Secondly, we investigate a channel tracking technique in OFDM systems in presence of both preamble and pilot symbols. In general, estimation accuracy is limited by the number of pilot subcarriers in each OFDM symbol. We propose an adaptively regularized LS estimation benefited by the initial estimated channel from preamble to improve estimation performance. The corresponding regularization parameter is adaptively determined by the accuracy of the latest estimate and the standard LS estimation. When pilot signals are insufficient to find the unique solution of the standard LS estimation, we propose a modified regularized LS estimation cooperated with DDCE to deal with the rank-deficient problem. The modified regularized LS estimation is charge of acquiring the up-to-date channel information for reducing error propagation in DDCE. On the other hand, the regularization matrix is modified as a diagonal one which is updated at the beginning of each frame by the new channel and noise statistic. Simulation results demonstrate the proposed method outperforms the standard LS with lesser pilots in low SNR environment. Simulation results show the proposed tracking technique outperforms the standard LS estimation and the traditional DDCE respectively.在本論文中，針對正交分頻多工 (orthogonal frequency division multiplexing, OFDM) 無線傳輸系統，我們提出了兩種有效改善效能通道估計演算法。 第一種方法主要是在具有傳輸分集(transmit diversity)的OFDM 系統中， 我們提出一個有效提升效能的低運算複雜度的通道估計法。在多傳輸天線的系統中，若能有效消除天線間干擾 (inter-antenna interference, IAI)，則通道估計就能簡化成如同在單輸入單輸出系統中所採用的方法。過去的研究中提出一個退耦 (decoupling) 的方法，藉由假設相鄰子載波的通道值為相同前提來有效的消除IAI。然而，但通道頻率選擇性增強時，殘餘的IAI會造成估計的準確度降低。於是，我們提出一個新的退耦方法，藉由簡化上述在頻域通道上的假設，並加入一個利用時間相關性的IAI補償方法，有效的將剩餘的IAI效應消除。 由過去所採用的退耦方法中，退耦過後的雜訊項會與傳送訊號產生相關性。當我們使用高維度的訊號調變時，由於其振幅不再為常數，則會使得上述的雜訊變異數在每個子載波上均不相同。因為此特性，則原本用來估計通道脈衝響應的最小平方 (Least Square, LS) 法不再是一個最佳解。於是，我們提出使用最佳線性無偏估計法 (best linear unbiased estimation, BLUE) 來取代原本的LS估計法，使得估計效能得以獲得改善。再進一步結合上述的退耦處理以及對於IAI的補償項，則通道估計的效能能夠大幅度的改善。在模擬結果中顯示，我們所提出方法的準確率，相當接近傳統同時估計所有傳送與接收間的通道的高計算複雜的方法。 在第二個通道估計法中，我們針對於通時具有訓練符元 (training symbol)以及引領 (pilot) 訊號的系統中，提出一個可適式的regularized LS通道估計演算法來有效的追蹤時變通道。在標準LS估計法中，通道的估計誤差受限於引領訊號的個數。為了能進一步提升LS估計法的效能，我們使用regularized LS來結合前一個時間的估計通道以及現在時間的LS估計結果，並以一個可調式的調整參數，來決定兩者在通道估計值中所佔的比例。此調整參數由兩者的估計誤差所決定，當其中一通道值較為不可靠時，則提高另一通道值所佔的比例，藉此有效的降低通道的估計誤差。 在引領訊號個數較通道階數為小的情況下，單純使用引領訊號無法找出通道的唯一解，上述的方法則無法使用。於是我們進一步結合直接決策通道估計法 (decision-directed channel estimation, DDCE) 來完成通道的追蹤。我們利用regularized LS估計法來得出具有更新資訊的通道值，並將此給於DDCE做直接的決策。不同於傳統DDCE利用前一個時間的估計通道估計值做直接決策，我們所提出的方法能夠有效的降低DDCE中錯誤傳遞的效應。在模擬結果中，我們所提出的regularized LS估計法，均能給夠給予標準的LS估計法以及DDCE估計法在效能上明顯的改善。誌謝....................................................................................................................................i 中文摘要...........................................................................................................................ii Abstract...........................................................................................................................iv Contents...........................................................................................................................vi List of Tables.................................................................................................................viii List of Figures.................................................................................................................ix 1. Introduction.................................................................................................................1 1.1 Background..............................................................................................................1 1.1.1 Channel Estimation in OFDM systems with transmit diversity...................3 1.1.2 Efficient Pilot-assisted Channel Estimation in OFDM systems...................6 1.2 Motivation and Contribution...................................................................................8 1.2.1 Improved Low-complexity Channel Estimation in OFDM Systems with transmit diversity..........................................................................................8 1.2.2 Adaptive Regularized LS Channel Estimation in Wireless OFDM Systems.........................................................................................................9 1.3 Dissertation Organization......................................................................................10 2. Improved Low-complexity channel estimation in OFDM systems with transmit diversity......................................................................................................................12 2.1 System Model........................................................................................................12 2.2 Literature Review..................................................................................................15 2.2.1 The Basic Channel Estimation...................................................................16 2.2.2 The Reduced-complexity Channel Estimation...........................................17 2.3 Proposed Channel Estimation Schemes................................................................21 2.3.1 Enhanced Channel Estimation by Removing the Residual IAI..................22 2.3.2 Improved Channel Estimation by Weakening the Equal-gain Assumptiom................................................................................................26 2.3.3 Improved Channel Estimation with Non-constant Envelope Modulation.................................................................................................32 2.4 Complexity Comparsion........................................................................................35 2.5 Simulation Results.................................................................................................41 2.5.1 MSE Comparison with Ideal Date..............................................................44 2.5.2 DDCE Performance Comparison...............................................................48 2.5.3 DDCE Performance with Non Constant Envelope Modulation.................51 2.6 Chapter Summary..................................................................................................54 3. Adaptively Regularized Least Square Channel Estimation in OFDM systems...55 3.1 System Model........................................................................................................55 3.2 Literature Review..................................................................................................57 3.2.1 Channel Tracking with Sufficient Pilot Signals.........................................57 3.2.2 Channel Tracking with Insufficient Pilot Signals.......................................60 3.3 Proposed Adaptively Regularized LS Channel Estimation...................................62 3.3.1 Adaptively Regularized LS Channel Estimation with Sufficient Pilots.....63 3.3.2 Adaptively Regularized LS Channel Estimation with Insufficient Pilots..67 3.4 Summary of the Adaptively Regularized LS Channel Estimation........................73 3.5 Simulation Results and Discuss.............................................................................75 3.5 Chapter Summary..................................................................................................86 4. Conclusion.................................................................................................................87 Bibliography...................................................................................................................8