Load pull compensation in multi user massive MIMO

本研究的目的是确定一种算法和/或过程，使其适合于负载拉动（相互耦合）的实际解决方案，这是“多用户大规模多进多出”（MIMO）系统中出现的问题。用于波束形成或MIMO的天线阵列要求天线元件在物理上彼此靠近。通过这种安排，有源天线元件会影响它们的邻居，反之亦然；修改天线阻抗（在给定频率下）和电磁场。这会导致辐射功率、射频波束方向和强度发生不必要的变化。数字预失真技术用于补偿功率放大器的非线性特性已得到广泛应用，也可用于补偿负载牵拉。由于天线单元之间的线性相位差均匀，例如在传统的波束转向中，数字预失真系数可以相对容易地计算出来。然而，在MIMO中，天线单元的相位和幅度配置是不规则的，因此天线单元之间的相互影响无法预测。动态MIMO信道矩阵，结合功率放大器传递函数和天线阵列特性，因此通过所述新算法进行处理，该算法计算每个单元的数字预失真系数，主动补偿每个单元的负载牵引力。本研究详细介绍了MIMO负载拉补偿算法和一个实用的实时硬件/固件解决方案。所得到的算法/实现在大频率范围内工作，适用于任何PA/天线组合。它也便宜，低功耗和可扩展到超大规模的MIMO。

The purpose of this research is to determine an algorithm and/or process that lends itself to a practical solution for load pull (mutual coupling), an issue that manifests itself in 'multi user massive multiple in multiple out' (MIMO) systems. Antenna arrays used for beam forming or MIMO require the antenna elements to be physically close to each other. With this arrangement, active antenna elements affect their neighbours and visa versa; modifying the antenna impedance (at a given frequency) and electromagnetic field(s). This causes unwanted shifts in radiated power, RF beam direction and intensity. Digital pre-distortion used to compensate for the non linear nature of power amplifiers has been used extensively and can also be employed to compensate for load pull. With a uniform linear phase difference across antenna elements, such as in conventional beam steering, the digital pre-distortion coefficients can be calculated relatively easily. However, with MIMO, the antenna elements phase and amplitude configuration is irregular and hence the effects of antenna elements on one another cannot be predicted. The dynamic MIMO channel matrix, in conjunction with the power amplifier transfer function and the antenna array characteristics are thus processed by said novel algorithm that calculates the digital pre-distortion coefficients for each element, actively compensating for load pull at each element. This research details the MIMO load pull compensation algorithm and a practical real-time hardware/firmware solution. The resulting algorithm/implementation works within a large frequency range and for any PA/antenna combination. It is also inexpensive, low power and scale-able to ultra massive MIMO.