在本文中，我們主要專注在巨量但有限(Large-but-Finite)的多輸入多輸出(MIMO)系統環境下，通道估測(channel estimate)中，使用不同的訓練序列(training sequence/pilot sequence)估測在比較貼近現實的系統(realistic system)中的通道，由於在基地台有大量天線的情況下，相鄰的基地台均使用相同的訓練序列會產生訓練序列汙染(pilot contamination)，此干擾可視為細胞間干擾(Inter-cell interference, ICI)，因此，本文希望可以藉由不同訓練序列方法，了解各序列對於訓練序列方法對於干擾程度的降低，並計算出訊號與干擾加雜訊比(SINR)，希望可以提升系統上行及下行的吞吐量(capacity)。 第一部分，會先介紹巨量天線多輸入多輸出系統(Massive MIMO system)，及本文使用大數量且有限天線之多輸入多輸出系統在較貼近現實的環境(realistic system)，及訓練序列汙染(pilot contamination)。第二部分，主要介紹不同的訓練序列方法Time shifted protocol在上述系統下可以有效的降低SINR，最後，模擬不同的訓練序列方法及預編碼(precoding)方法在較現實的系統中，下行的總吞吐量(total capacity)比較。 ;In this paper, we focus on forward and reverse transmission of non-cooperative multi-cellular time-division duplexing (TDD), large-but-finite MIMO system that the number N of antennas per base station (BS) are large (more than 100) and the number of user terminals per cell are also . This system we consider channel estimation, pilot contamination, path loss and antenna correlation for each link. We simulate achievable rates with beamforming (BF)/regularized zero-forcing (RZF) linear precoders and minimum-mean-square -error(MMSE) on uncorrelated channels from previous work and more realistic system. Through raising signal-to-interference plus noise ratio (SINR) that is limited by pilot contamination, we derive better performance. The other we consider using time-shifted protocol to compare with the previous simulation. We derived that through time-shifted protocol to grouping the cells can reduce the inter-cell interference caused by pilot contamination. At the beginning, we introduce MIMO system and Massive MIMO system. Second, we described large-but-finite MIMO system which is the system we simulation. Finally, using the methods mentioned above, deriving simulation results.
