在混合式波束成型下通道估測之基頻預編碼器和組合 器設計

Abstract

近年來，混合式波束成型架構被提出來解決毫米波通訊的嚴重路徑耗損和硬體限制問題，在完美通道資訊被給定的假設下，我們可以設計混合式預編碼器和結合器，接著在毫米波通道中傳輸資料，然而，若我們無法得知通道資訊，我們要事先估測它，所以，一些通道估側的方法已被提出，其中之一是開迴路通道估測，它利用了毫米波通道的稀疏性質且使用壓縮感知的技術估測出這個稀疏的通道，我們的目標是藉由增加基頻預編碼器和結合器的設計，改良此開迴路通道估測的方法，更精確地來說，在增加了基頻的設計後，我們可以得到在壓縮感知技術中較低的感測矩陣相干性，來獲得更好的估測效能，我們應用一個來自於數學相關文獻的演算法到我們提出的工程問題上，我們的模擬結果顯示兩點，第一點是我們提出的設計相較於之前開迴路通道估測，能得到更低的感測矩陣之相干性，第二點是感測矩陣的低相干性對於通道估測的效能有正面的影響，所以，我們提出的設計有著比先前開迴路通道估測的方法更低的通道估測誤差。

In recent years, the hybrid beamforming architecture has been proposed to deal with the severe path loss and the hardware limitation of millimeter wave (mmWave) communication. Under the assumption that the perfect channel state information (CSI) is known, we can design hybrid precoder and combiner and then transmit data in the mmWave channel. However, if CSI is unknown, we need to estimate it in advance. Therefore, some channel estimation methods have been proposed. One of them is the open-loop channel estimation, which utilizes the sparsity of mmWave channel and estimates the sparse channel by the compressed sensing technique. Our objective is to improve the open-loop channel estimation method by adding the baseband design. To be more specific, after adding baseband design, we can get lower mutual coherence of the sensing matrix in the compressed sensing technique, in order to reduce the channel estimation error. We apply an algorithm from mathematical literature to our proposed engineering problem. Our simulation results indicate two points. One is that our proposed design could get lower mutual coherence of the sensing matrix than the previous open-loop channel estimation. The other is that the low mutual coherence property has a positive effect on the channel estimation performance, so our proposed design has lower channel estimation error than the previous open-loop channel estimation method.