毫米波 (Millimeter wave) 系統被視為未來無線通訊協定的重要應用，可以提供每秒十億位元的高資料速率傳輸性能，然而在此系統的高頻段中將引入巨大的傳輸損耗使得訊號品質大幅下降，則波束成型 (Beamforming) 被視為毫米波系統不可或缺的關鍵技術。此外，無線通訊會受到環境的影響產生多路徑傳播效應，尤當接收端屬移動目標源更會使通道衍生時變現象並迫使波束校準的困難度增加，因此訊號的出發角 (Angle of Departure, AoD) 與入射角 (Angle of Arrival, AoA) 的估計與追蹤儼然成為毫米波系統的核心議題。本論文考量一多輸入多輸出 (Multiple-Input Multiple Output, MIMO) 的均勻線性陣列 (Uniform Linear Array, ULA) 的天線架構，以分時 (Time division) 波束搜索 (Beam searching) 的方式對空間進行初始的探測，進而透過輔助波束對 (Auxiliary BeamPair, ABP) 的數值技巧，根據訊號強度做多路徑的檢測並估計角度資訊。而為克服時變現象引發傳送及接收端的波束失準或不匹 配，我們以區塊衰弱通道的環境為基礎並提出無跡卡爾曼濾波器(Unscented Kalman filter) 自適應演算法逐每個傳輸資源塊進行高效 率的波束追蹤 (Beam tracking)，並利用模擬結果進行性能分析與討 論。 ;The millimeter wave (mmWave) communication is deemed as a promising technology for the approaching generation of wireless com munications, which can provide high data rate transmission over multi-gigabit per second. However, the mmWave spectrum introduces high propagation attenuation that significantly degrades signal quality, and thus, the beamforming technology becomes indispensable for a mmWave system. Moreover, there are several uncertainties in wireless communi cations like the multipath propagation effect and especially, the moving mobile equipment generates fast time-varying channels and makes the beam alignment more inconvenient in the real time environment. Therefore, estimation and tracking of angle-of-departure (AoD) and angle-of arrival (AoA) are obviously the major issue for a mmWave system. This thesis considers a multiple antennas system based on a uniform linear array (ULA) structure. We propose a beam search procedure in a time di vision manner for initialization, utilizing the auxiliary beam pair (ABP) nu-merical technique to perform multipath detection and estimating an gle information based on signal strength. Furthermore, to overcome the beam misalignment or mismatch resulted from the time-varying channel, we employ the unscented Kalman filter to adaptively perform efficient beam tracking over a block fading channel. Finally, we demonstrate the proposed system with computer simulation results and performance analysis in a typical mmWave environment.
