本文考慮在大規模多重輸入多重輸出(massive multiple-input multiple-output, massive MIMO)分時雙工(time-division duplex ,TDD)多細胞群組(multi-cell grouping)系統下的引示訊號汙染(pilot contamination)預編碼器(precoding)設計。在下世代無線通訊中為了達到高傳輸速率,Massive MIMO及雙層多細胞系統(two-layer multi-cell system)架構已被視為5G通訊系統的潛力股,雙層多細胞系統此架構分成大細胞與小細胞,大細胞為現行的4G架構,小細胞則為5G高傳輸量的架構,為了提高傳輸量需要準確的通道狀態資訊(channel state information, CSI),由於在多小細胞之間重複使用相同的引示訊號序列(pilot sequence),在基地台獲得的通道估計(channel estimation)不僅包含目標使用者的通道資訊,而且還包含來自其他小細胞之使用者的通道資訊,使得基地台通道估計不準確,造成引示訊號汙染,因此開發出了基於多細胞的引示訊號汙染預編碼,研究兩種引示訊號汙染預編碼對抑制污染後系統總和傳輸率(sum rate)的效果,並且透過多細胞群組(grouping)架構,對多細胞系統分配使用不同的引示訊號,消除相鄰細胞的引示訊號汙染,降低單獨使用預編碼時的高系統預算複雜度。 In the thesis, a pilot contamination precoding design is proposed based on massive multiple-input multiple-output (MIMO) time-division duplex (TDD) multi-cell group systems. In order to achieve high transmission rate, the integration of massive MIMO and two-layer multi-cell system architecture has been regarded as a promising technology in future fifth-generation (5G) communication system. The two-layer multi-cell architecture consists of macro cells and small cells. Macro cells use the existing 4G system while small cells use 5G technology to provide higher high data rate and system capacity. In order to improve data rate of wireless networks is required with accurate channel states information. Since multi-cell use the same pilot sequences for channel training, the channel estimation is to be inaccurate and cause pilot contamination. Therefore, pilot contamination precoding is designed based on multi-cell system, and study the effects of two pilot contamination precoding designs in terms of system sum-rate. Through the multi-cell group architecture assigned to use different pilot sequence to small cells, eliminate the pilot contamination of adjacent cells and reduce the high system complexity budgets when using pilot contamination precoding alone.