| 摘要: | 相比於前面幾代的通訊系統，5G 對通訊系統提出更高的要求、更快的傳輸速度、更低的資料延遲以及更好的效率。這些要求都是得益於對毫米波 (Millimeter wave, mmWave) 的應用與開發。但在另一方面，巨大的傳輸損耗使通訊品質大幅下降。波束成型 (Beamforming) 為克服上述影響的關鍵技術。在混合波束成形架構下，射頻鏈路的數量決定了成本的高低，射頻鏈路透過相移器連接到每一支天線上，導致提高了硬體設計上的複雜度，所以部分連結的結構就被提出，在射頻鏈路只連結到部分天線上，並沒有完全連接到所有天線上，這個架構就被稱作混合波束部分連結架構，這對昂貴的毫米波硬體來說，它在設計上的複雜度有所降低，且更具實際上的應用價值。可由於射頻鏈路沒有連接到所有天線上，這導致頻譜效率變低，從而資料傳輸的準確率下降，這是硬體上的限制。在本文中，我們採用多路徑三維通道(multi-path three-dimensional (3D) channel) 為通道環境，並提出了結合部分連結的波束成型演算法與幾何平均分解 (GMD) 的算法，透過演算法得出它的類比預編碼器與結合器，再使用幾何平均分解去優化數位預編碼器與結合器的部分，它可以避免複雜的比特分配。與傳統的基於 SVD 的混合預編碼不同，它可以將毫米波大規模 MIMO 通道轉換為具有相同增益的子通道，因此不再需要額外的功率分配，使得調變和編碼的設計複雜度降低，讓部分連結系統可以用低複雜度實現更好的錯誤率 (BER) 性能，此外我們一樣是討論錯誤率得情況下，將部分連結天線陣列優化方式，套用到智能反射板上，看是否能夠跟天線陣列的訊號處理一樣有同樣的
效果，最後利用模擬結果進行性能分析與討論。;Compared with previous generations of communication sys tems, 5G puts forward higher requirements, faster speed, lower delay and better efficiency for communication systems. These requirements are all benefited from the application and development of Millimeter wave. But on the other hand, the huge transmission loss makes the com munication quality drop significantly. Beamforming is a key technol ogy to overcome the above-mentioned effects.In the hybrid beamform ing architecture, the number of Radio Frequency chain determines the cost. The Radio Frequency chain are connected to each antenna through a phase shifter, which increases the complexity of hardware design, so the structure of partial connection is proposed that when the Radio Fre quency chain is only connected to some antennas, but not fully connected
to all antennas, this architecture is called hybrid beamforming partial connection architecture. However, since the Radio Frequency chain is not connected to all antennas, the spectrum efficiency becomes low, and the accuracy of data transmission decreases. This is a hardware limita tion. In this paper, we use a multi-path three-dimensional (3D) chan nel as the channel environment, and propose an algorithm combining partially connected beamforming algorithm and geometric mean decom position(GMD), through the algorithm get its analog precoder and com biner, and then use GMD to optimize the part of the digital precoder and combiner, it can avoid complex bit allocation. Different from the tradi tional SVD-based hybrid precoding, it can convert the mmWave massive MIMO channel into the same sub-channel, so no additional allocation is needed, the design complexity of modulation and coding is reduced, and the partial connection system’s BER performance can be achieved with low complexity. In addition, we are also discussing the BER, and apply partial connected array optimization method connected to the RIS to see if it can have the same effect as the signal processing of the antenna array, and finally the simulation results are used for performance analysis and discussion. |
