Iterative carrier frequency offset/channel estimation and robust detection for MC-CDMA systems

[[abstract]]在多載波分碼多重存取(multi-carrier - code division multiple access，MC-CDMA)系統中，採用子載波展頻傳送資料的形式可以提供較大的頻率分集，而且在頻率選擇性衰減通道中，增加抗衰減(fading)的能力。但是此系統容易受到載波頻率偏移(carrier frequency offset，CFO)破壞子載波正交性，以及需要知道通道狀態資訊(channel state information，CSI)才能做偵測(detection)。本論文第一部分，在MC-CDMA系統，針對舊有的通道估測方法與載波頻率偏移估測方法提出一個疊代式(iterative)的估測方法，並模擬疊代式估測方法的效能。結果顯示此方法可以同時估測通道狀態資訊與載波頻率偏移之外，也能有效地降低運算複雜度，從而獲得不錯的效能。從本論文第一部分可得知每一級(stage)均有通道估測誤差值與載波頻率偏移估測誤差值，基於這，我們假定在一定範圍內的通道誤差與CFO誤差，並以最大化訊號對干擾及雜訊比(signal-to-interference-plus-noise ratio，SINR)的前提下，常用之最小變異(minimum variance，MV)偵測器與對角載入最小變異(diagonal loading minimum variance，DLMV)偵測器都會造成效能大幅下降。因此在論文第二部分，在假定的通道狀態資訊與載波頻率偏移估測下，透過使用二階錐形問題(second-order cone programming，SOCP)，我們提出強健性(robust)偵測器，可以改善MC-CDMA偵測時的效能。利用不同的通道估測誤差值與載波頻率偏移估測誤差值，從模擬圖可以觀察到，我們所提出的強健性偵測器效能會比傳統的最小變異與對角載入最小變異偵測器來的好。[[abstract]]Multi-carrier code division multiple access (MC-CDMA) system utilizes multiple subcarriers for data transmission, which provides frequency diversity and enhances the capability of mitigating fading in frequency selective fading channels. However, the effect of carrier frequency offset (CFO) may destroy the subcarrier orthogonality. In addition, the channel state information (CSI) is required for data detection. In the first part of the thesis, based on the conventional channel and CFO estimation methods, we propose a new scheme for estimating the channel and CFO iteratively. Its performance is fully analyzed. Simulation results show that the proposed scheme can estimate both channel and CFO effectively, and reduce the computational complexity at the same time. From this part, we observe that some errors exist certainly during the estimation of both channel and CFO. In this case, the performance of the conventional minimum variance (MV) and diagonal loading minimum variance (DLMV) detection based on maximizing the signal-to-interference-plus-noise ratio (SINR) degrades significantly. In the second part of the thesis, assuming that the estimation errors are within a certain range, we propose a robust detector solved via the second-order cone programming (SOCP). It improves the detection performance of the MC-CDMA system under these scenarios. Simulation results show that our robust detector can outperform the conventional MV and DLMV detectors with different channel and CFO estimation errors.[[note]]碩