於本論文我們討論上行的多使用者(Multiuser,MU)多重輸入多重輸出(Multiple Input Multiple Output,MIMO)在多載波系統伴隨通道編碼(Channel Coding)系統之效能,以探討空間分集(Space Diversity)和頻率分集(Frequency Diversity)對系統之影響。我們知道基於子載波訊號處理方法雖然可以得到最佳的系統效能,但亦伴隨著很高的複雜度。因此,近代有研究提出基於符元處理之技術,即在傳送端/接收端分別只需要使用一個反離散傅立葉/傅立葉轉換,從而大幅降低複雜度。傳統基於符元處理之準則主要為最大化所有子載波之訊號訊雜比總和,此方法之缺點為子載波的訊號訊雜比常常無法平衡,造成整個系統的效能下降。在基於符元處理前提之下,我們提出使用最大-最小公平準則(Max-Min Fair Criterion)以解決此問題,並利用伴隨較低複雜度的正半定放寬(Semidefinite Relaxation,SDR)求解。在論文第一部份,我們於多使用者多重輸入多重輸出正交分頻多工存取系統下考慮已知通道狀態資訊(Channel State Information,CSI)之情況,並在傳送端進行傳輸波束形成。在論文第二部份,為了解決正交分頻多工存取所造成的高峰值平均功率比(Peak-to-Average Power Ratio, PAPR),我們提出在單載波正交分頻多工存取系統(Single-carrier Frequency-Division Multiple Access,SC-FDMA)使用最大-最小準則。我們將於不同之訊號處理準則和環境下模擬結果。
In this thesis, we discuss the performance of uplink multiuser (MU) multiple input multiple output (MIMO) multicarrier systems with channel coding. The effect of space and frequency diversities is then studied. In this aspect, it is known that subcarrier-based signal processing techniques can provide the best performance. However, the computational complexity will be prohibitively high. Therefore, symbol-based processing techniques have recently been proposed for complexity reduction. Only one inverse discrete fourier transform (IDFT)/DFT block is required at the transmitter/receiver. Conventional symbol-based processing is to maximize the sum of the signal-to-noise ratio (SNR) values achieved in all subcarriers. The disadvantage of it is that the SNR performance of different subcarriers is often not balanced. This will degrade the whole system performance. Here, to solve this problem, we propose to use the max-min fair criterion for symbol-based processing. Semidefinite relaxation (SDR) is used to approximate the solution with low complexity. In the first part of the thesis, we consider the scenario in which channel state information (CSI) is known at the transmitter, and then transmit beamforming is employed for the MU-MIMO-OFDMA systems. In the second part of the thesis, to solve the problem of high Peak-to-Average Power Ratio (PAPR), we propose to use the max-min fair criterion for symbol-based processing in MU-MIMO-SC-FDMA(Single-carrier Frequency-Division Multiple Access,SC-FDMA) systems. Finally, simulation results show the performance of difference schemes with difference environment settings.