| 摘要: | 具有豐富頻譜資源的毫米波 (Millimeter wave, mmWave)
通訊系統被視為下一時代無線通訊系統的潛力技術，在高頻
段傳輸可獲得每秒十億位元的高資料速率，但同時也引入了
巨大的傳輸損耗使通訊品質大幅下降，為了確保良好的通訊
品質，高指向性的波束成型 (Beamforming) 被視為毫米波通訊
系統中不可或缺的關鍵技術，因此精準的訊號出發角 (Angle
of Departure, AoD)、入射角 (Angle of Arrival, AoA) 及路徑增益顯得特別重要。特別是移動通訊 (mobile communications) 場景，環
境的些許變化導致傳送端與接收端的波束錯位，使接收訊號
的品質明顯下降，因此角度與路徑增益的估計與追蹤成為毫米波通訊系統的核心研究主題。本論文考慮單一使用者多輸
入多輸出正交分頻多工 (Multiple-Input Multiple-Output Orthogonal
Frequency-Division Multiplexing , MIMO-OFDM) 的均勻線性陣列
(Uniform Linear Array, ULA) 天線架構，並假設傳送端與接收端使
用全連接混合波束成型 (full connection hybrid beamforming) 架構，
在初始階段，利用分層波束訓練 (hierarchical beam training) 對空間
進行粗掃描 (coarse search) 及細掃描 (fine search)，使獲得最佳的
波束匹配，並假設傳送端與接收端的波束中心角為初始估計的訊
號出發角、入射角，而初始路徑增益由最小平方法 (least squares)
求出，接著利用正交匹配追蹤 (Orthogonal Matching Pursuit, OMP)
取得混合波束成型架構之預編碼器 (precoder) 與結合器 (combiner)
權重。我們採用多路徑二維通道 (multi-path two-dimensional (2D)
channel) 為通道環境，並提出基於決策反饋 (Decision Feedback) 無
跡卡爾曼濾波器 (Unscented Kalman Filter) 自適應演算法，不需要
UKF 波束追蹤前導序列 (preamble) 也能克服時變的傳送端與接收
端波束匹配，達到高效率低時間成本的波束追蹤 (beam tracking)，
並利用模擬結果進行性能分析與討論。;Millimeter-Wave (mmWave) communication system with abun dant spectrum resources is regarded as the potential technology of the
next-generation wireless communication system. The high data rate of
one gigabit per second can be obtained in high-frequency band trans mission, but it also introduces a huge transmission loss that greatly re duces communication quality. In order to ensure good communication
quality, beamforming with high directivity is regarded as an indispens able key technology in the mmWave communication system. Therefore,
the precise signal Angle of Departure ( AoD), Angle of Arrival (AoA),
and path gain are critical. Especially in mobile communications scenar ios, slight changes in the environment cause the beams at the transmitter and receiver to be misaligned, which significantly degrades the quality of
the received signal. Therefore, the estimation and tracking of angle and
path gains have become the core research topics of the mmWave com munication systems. This paper considers a single-user Multiple-Input with high efficiency and low time overhead. finally, the simulation re sults are used for performance analysis and discussion.
Multiple-Output Orthogonal Frequency-Division Multiplexing (MIMO OFDM) Uniform Linear Array (ULA) antenna architecture. It assumes
that the full connection hybrid beamforming architecture is used for both
the transmitter and receiver. In the initial stage, hierarchical beam train ing is used to perform coarse search and fine search on the angle space, to
obtain the best beam matching, and to assume that the beam center angles
of the transmitter and receiver end are the initially estimated signal AoD
and AoA, respectively, and the initial path gain is obtained by the Least
Squares (LS) method, Orthogonal Matching Pursuit(OMP) obtains the
precoder and combiner weights of the hybrid beamforming architecture.
We use a multi-path two-dimensional (2D) channel as the channel en vironment, and propose an adaptive algorithm for the decision feedback
based Unscented Kalman Filter(UKF), which does not require UKF beam
tracking preambles and can also overcome the time-varying beam match ing between the transmitter and the receiver, and achieve beam tracking |
