To compensate the severe path loss of millimeter-wave channel, large-scale antenna arrays are required to provide beamforming gains for millimeter wave communication systems. Since the millimeter wave (mmWave) MIMO channel has strong directional selectivity, it is necessary for transmitter and receiver to perform beam alignment for obtaining maximum beamforming gains. In this paper, we first propose an adaptive beam alignment scheme. Based on the proposed scheme, adaptive hierarchical traversal beam search algorithm and adaptive hierarchical iterative beam search algorithm are proposed for beam alignment. Both of the algorithms adjust the training time adaptively by exploiting the beam detection algorithm to improve the success rate of beam alignment and reduce beam training overhead. Furthermore, in order to improve the transmission data rate and the beam alignment reliability of the system, maximum transmission data rate is adopted as the termination criterion of beam training. Simulation results show that the proposed algorithms have higher transmission data rate and beam alignment success rate compared with the existing beam search algorithms under both line-of-sight (LOS) and non-line-of-sight (NLOS) channels. Moreover, the proposed algorithms are more reliable to dynamic scenarios, especially the short distance between transmitter and receiver.

为了补偿毫米波信道的严重路径损耗，需要大型天线阵列为毫米波通信系统提供波束成形增益。由于毫米波（mmWave）MIMO信道具有很强的方向选择性，因此发射器和接收器有必要执行波束对准以获得最大的波束成形增益。在本文中，我们首先提出了一种自适应光束对准方案。基于所提出的方案，提出了用于波束对准的自适应分层遍历波束搜索算法和自适应分层迭代波束搜索算法。两种算法都通过利用波束检测算法来自适应地调整训练时间，以提高波束对准的成功率并减少波束训练开销。此外，为了提高传输数据速率和系统的波束对准可靠性，采用最大传输数据速率作为波束训练的终止准则。仿真结果表明，与现有的波束搜索算法相比，该算法在视距（LOS）和非视距（NLOS）信道下均具有更高的传输数据速率和波束对准成功率。此外，所提出的算法对于动态场景尤其是发射机与接收机之间的短距离更可靠。仿真结果表明，与现有的波束搜索算法相比，该算法在视距（LOS）和非视距（NLOS）信道下具有更高的传输数据速率和波束对准成功率。此外，所提出的算法对于动态场景尤其是发射机与接收机之间的短距离更可靠。仿真结果表明，与现有的波束搜索算法相比，该算法在视距（LOS）和非视距（NLOS）信道下均具有更高的传输数据速率和波束对准成功率。此外，所提出的算法对于动态场景尤其是发射机与接收机之间的短距离更可靠。