|Abstract: ||多輸入多輸出(Multiple Input Multiple Output, MIMO)技術可提高無線通訊系統資料傳輸量；正交分頻多工調變(Orthogonal Frequency Division Modulation, OFDM)技術具有高頻譜效率和對抗多重路徑衰減通道的效應；而空時碼(Space-Time Coding, STC)則利用空間分集特性編碼來增加傳送資料量。因此，結合STC和MIMO-OFDM技術已成為第四代寬頻行動通訊的標準技術之ㄧ。本論文針對系統架構為二對二天線的環境，在MIMO STC-OFDM無線傳輸系統進行等化器研究。 首先，利用估測之通道參數發展最小平方誤差(Least Square Error, LSE)和最小均方誤差(Minimum Mean Square Error, MMSE)條件下的LS濾波器和Wiener濾波器以解調傳送訊號。為提高等化器效能，依據STC-OFDM訊號特性，論文中提出同時使用正向及反向接收資料來設計等化器。此法不需額外增加天線數目或者要求傳輸端再傳送一次資料，就能多出一倍的資料量，相對於僅用正向資料設計的等化器，將有效於降低解調端的位元錯誤率。 接著，為因應時變的通訊環境以及減緩通道估測誤差所造成的接收效能衰減，研究直接於接收端設計適應性等化器。此等化器以RLS (Recursive Least Square)法進行權重調整，論文中分別針對僅使用正向資料和同時使用正反向資料提出正向RLS和正反向RLS法。模擬中證實正反向RLS法所設計的等化器在收斂速度和接收效能均優於正向RLS法。此外，因盲蔽式等化器的效能取決於決策函數參數的選取，論文中提出一適應性決策函數參數調整法，以提升時變通道環境下，盲蔽式等化器的接收效能。|
Abstract Multiple-input multiple-output (MIMO) technique can improve the data rate of wireless communication. Orthogonal Frequency Division Modulation (OFDM) technique can alleviate the performance degradation due to multipath interference. Space-Time Code (STC) uses the diversity in space and time domains to increase the transmission data rate. Therefore, the combination of MIMO, STC and OFDM techniques, called the MIMO STC-OFDM system, has become one of standards in the fourth-generation (4G) mobile wideband communication. In the study, we proposed adaptive equalizer based on recursive least square (RLS) algorithm. First, we use the parameters of channel estimation to develop LS filter and Wiener filter under the condition of LSE and MMSE to demodulate the transmitted data. To improve the accuracy of equalizers, we combine the forward and backward received signals to design the winner filter by utilizing the symmetric property of MIMO-STC-OFDM data. Comparing the equalizer with forward-only data, our proposed equalizer with forward and backward data can reduce the bit error rate in the receiver efficiently. Next, in order to apply to nonstationary communication environment and reduce the performance degrading cause by estimation errors, we designed an adaptive equalizer base on Recursive Least Square (RLS) algorithm. In the study we developed blind equalizers based on the forward-only RLS and forward-backward RLS algorithms, which using the forward-only and the combined forward and backward receiving signals, respectively. Moreover, since the performance of blind equalizer depends on the decision function, we further designed the blind equalizers based on an adaptive decision-directed algorithm. Simulations were conduced based on the BPSK signals, and validated that the proposed adaptive equalizer combined forward and backward signals perform better than the equalizer with forward-only signals. Besides, our developed adaptive decision-directed blind equalizer can demodulate the receiving signals immediately in the nonstationary environment.