近年來由於無線通訊的蓬勃發展,使用多輸入多輸出(MIMO)系統已經成為無線通訊發展的主流趨勢,此外,MIMO衍伸出許多相關技術,其中預編碼(precoding)技術是本論文所探討的重點。一般來說,在實際的系統中,利用常見的挑選法則(selection criterion)從編碼簿(codebook)中可以挑選出最適合的碼字(codeword)當作預編碼矩陣(precoding matrix),本論文所使用的編碼簿為LTE-Advanced下行鏈路的編碼簿。但是,隨著接收天線數目與傳送的數據串流(spatial streams)數目越多,則挑選法則所耗費的計算複雜度就會越大。為了能夠在實際的系統中實現,本論文選擇了四個複雜度較低的挑選法則,分別為MMSE Geometric Mean Criterion、MMSE Trace Criterion、Max Post Processing SNR Criterion以及Post MMSE Criterion,而四個挑選法則中,錯誤率最低的是MMSE Trace Criterion,而複雜度最低的則是Max Post Processing SNR Criterion,而它的錯誤率是最高的,透過電腦模擬在不同接收天線數與傳送數據串流數情況下,比較它們的效能與複雜度分析。
As the wireless communications prevails, the various MIMO techniques attract much attention in recent years. Among the various MIMO techniques, the limited feedback precoding is the focus of this thesis. The precoding matrix at the transmitter side is selected from a finite codebook. The LTE-Advanced downlink codebook is considered in this thesis. As the number of receiver antennas or spatial streams grow, the computational complexity of selection criteria also grow. In order to implement to practical system, we use four practical selection criteria of low computational complexity of MMSE Geometric Mean Criterion, MMSE Trace Criterion, Max Post Processing SNR Criterion, and Post MMSE Criterion respectively. In these criteria, the lowest BER criterion is the MMSE Trace Criterion. The lowest computational complexity is the Max Post Processing SNR Criterion, but its BER is the highest. We compare the performance and computational complexity of selection criteria with different receiver antennas and spatial streams.