Because millimeter wave (mmWave) systems can span notably wide spectral bands, mmWave systems are expected to dominate fifth-generation (5G) communication systems. Due to the short wave-length of mmWave radiation, multiple-input multiple-output (MIMO) systems can use massive antennas and precoding technology to overcome signal attenuation in mmWave channels. However, the cost and power consumption of radio frequency (RF) chains would increase substantially with the number of antennas. Hence, hybrid beamforming was proposed to reduce the number of RF chains in massive MIMO systems. Hybrid beamforming involves RF beamforming matrix construction and baseband precoding matrix derivation. This study focused on the design and implementation of an algorithm for the RF beamforming matrix construction for mobile environments. Accordingly, this study presents a mixture particle filter that exploits the temporal continuity of beam clusters in a mobile mmWave channel to reduce the computational complexity of RF beamforming matrix construction. Moreover, this beam-tracking particle filter is based on parallel processing architecture to support the tracking of multiple beam clusters in the mmWave channel. Finally, the beam-tracking particle filter was implemented on a field-programmable gate array platform and was verified in a hybrid beamforming system for mmWave MIMO systems. The particle filter processor achieved a maximal throughput of 9.198k matrices/s with a clock rate of 192 MHz, which could support a speed of up to 88.5 km/h for mobile users.

由于毫米波（mmWave）系统可以显着跨越很宽的光谱带，因此毫米波系统有望在第五代（5G）通信系统中占主导地位。由于毫米波辐射的波长短，多输入多输出（MIMO）系统可以使用大型天线和预编码技术来克服毫米波信道中的信号衰减。但是，射频（RF）链的成本和功耗将随着天线数量的增加而大大增加。因此，提出了混合波束成形以减少大规模MIMO系统中RF链的数量。混合波束成形涉及RF波束成形矩阵构建和基带预编码矩阵推导。这项研究集中于移动环境中射频波束成形矩阵构建算法的设计和实现。因此，这项研究提出了一种混合粒子滤波器，该滤波器利用移动毫米波通道中波束簇的时间连续性来降低RF波束形成矩阵构建的计算复杂性。此外，该束跟踪粒子滤波器基于并行处理架构，以支持mmWave通道中多个束簇的跟踪。最后，束跟踪粒子滤波器在现场可编程门阵列平台上实现，并在用于mmWave MIMO系统的混合波束形成系统中得到验证。粒子滤波器处理器以192 MHz的时钟速率实现了最大9.198k矩阵/ s的吞吐量，这对于移动用户而言可以支持高达88.5 km / h的速度。