適用於毫米波雙重極化陣列天線之混合式波束成型與多使用者前置編碼技術

Abstract

此博士論文針對毫米波頻段上的無線通訊系統，提出在雙重極化平面陣列天線上能混合使用射頻信號處理與基頻信號處理的架構。在此架構之下，吾人探討其所形成之雙重極化混合式多輸入輸出系統被應用於毫米波頻段的無線通訊時，其通道容量的分析和通訊演算法的設計。此外，由於此博士論文所提出的通訊演算法需要傳送端知道通道的資訊，在此情況下，通道的資訊會由接收端估計出來並廻授至傳送端。因此，吾人進一步地去研究在更廣汎的時變平坦衰減多輸入輸輸出通道下，利用訓練符元估計通道的傳輸機制其傳輸率和估測通道資訊的平均方根錯誤之間的關係。 在雙重極化混合式多輸入輸出系統的通道容量分析中，吾人考慮不同的混合式架構可得到的通道容量，這些架構中有可利用多輸入輸出通道的多樣性來獲得空間多工增益的功能，也有可利用波束成型所造成的指向性來加強陣列增益的功能，甚至有些架構可同時使用以上兩種功能。在演算法的設計中，吾人考慮在室內和室外環境下，透過雙重極化混合式多輸入輸出系統在毫米波無線電的多用戶傳輸的情境。在室內空間以空間分割多重存取方式傳輸多媒體資料流時，為了能在無法預知的多重存取干擾和符元間干擾存在情況下保持信號的品質，吾人在接收端只能回傳有限的資訊和傳送端知道各用戶的位置的情況下，設計了在傳送端使用的穩健式混合式波束成型演算法，此演算法以最小的傳送功率來保持各用戶資料流的信號干擾加雜訊比。此外，考量到波束成型的效能對於陣列天線中相位移相器的相位誤差極為敏感，兩種穩健式波混合式波束成型被提出同時抵抗多重存取干擾、符元間干擾存和相位誤差。另一方面，為了在室外蜂巢式毫米波無線通訊中保證各用戶在下行通訊機制的通訊品質，在基地台完美地知道通道資訊且傳送功率被限制的情況下，一個同時設計射頻波束成型器、基頻前置編碼器、基頻等化器的演算法被提出。最後，在時變平坦衰減通道下，對於一個使用訓練符元進行通道估的通訊系統，吾人以克勞姆-羅下限的角度去分析其估測通道資訊的平均方根錯誤對於系統的可得傳輸率的影響。藉由此論文的結果，可以幫助我們了解何種混合式架構適合於使用雙重極化平面陣列的毫米波無線電，以及了解在此情況下混合式架構可得到效能和正確的使用方式。

A cost-effective architecture is proposed for a joint radio frequency (RF) and baseband (BB) hybrid signal processing method that uses dual-polarized (DP) planar antenna arrays (PAA) in millimeter wave (mmWave) radio transmissions. Under the DP hybrid architecture, the channel capacity and the transmission algorithms of the corresponding DP hybrid multiple input multiple output (MIMO) systems in mmWave frequency bands are studied. According to the fact that the channel state information (CSI) typically estimated at the receiver is needed at the transmit side for operating the transmission algorithm designed in this dissertation, we further investigate the influence of channel estimation on the achievable data rate in more general time-varying flat faded MIMO channels when the training-based transmission scheme is applied. The channel capacity of the indoor mmWave transmission is studied for different configurations of the DP hybrid architecture which can exploit the spatial multiplexing gain of MIMO channel diversity, or enhance the array gain with beamforming (BF) directivity. The algorithms for the multi-user transmission through mmWave radio are considered in both indoor and outdoor environments. To maintain signal qualities under unpredictable multiple access and inter-symbol interferences (MAI/ISI) encountered in spatial division multiple access (SDMA) in wireless multimedia streaming to which indoor mmWave mainly applies, transmit robust beamformers are designed to maintain the signal to interference-plus-noise ratio (SINR) for each user with the minimum total transmit power, limited feedbacks and the prior knowledge of users' locations. In addition, as the performance of beamforming in SINR is sensitive to uncertainties in the phase shifters of PAA, two kinds of robust formulation are also proposed to jointly combat the MAI, ISI and phase uncertainties. On the other hand, to guarantee the quality of service of each user under the limited power consumption in outdoor cellular mmWave radio, a joint design of RF and BB signal processing in hybrid architectures is proposed for the downlink scenario with the assumption that the perfect CSI of the RF propagating channel is known at the base station. Finally, for training-based MIMO systems in time-varying flat faded channels, we discuss the achievable rate under the influence of channel estimation which is examined from the perspective of Bayesian Cramér-Rao lower bound. The results in the dissertation help configure hybrid architecture more practical and provide the guideline and the achievable performance of using hybrid architecture with DP-PAA in mmWave radio transmissions.