Power Combining Networks for MIMO Transmission

本文研究了功率组合在多天线无线电中的应用。现代无线电系统通常使用多个发射天线和适当的信号编码来提高无线链路的吞吐量和可靠性。理想情况下，每个放大器的额定功率应达到所有天线传输的最大总功率（或总和）。出于成本和效率的考虑，额定功率通常被设计得较低（通常是最大总功率除以天线数量），这可能导致显著的性能损失。本研究探索了一种使用功率组合网络（PCN）的解决方案，该解决方案使放大器的输出能够切换/组合到任何天线中，从而将功率从较弱的天线转移到较强的天线，从而提高链路性能。MIMO（多输入多输出）天线发射机的最大输出功率是有限的，由于生物安全的考虑。当每个功率放大器的额定功率达到系统的全部发射功率时，MIMO传输性能最佳。如果出于成本和效率方面的考虑，放大器的额定功率较小，那么在信道不平衡的情况下，M天线的最坏情况损失可能达到log M dB。PCN可以显著降低这种损耗，以阵列增益的损失换取更好的信道耦合效率。PCN的使用改变了有效的MIMO信道，因此也可以改变最佳传输编码。本研究的目标是研究PCN的快速选择算法、合适的信号编码方案和最佳的可实现性能。

This research studies the use of power combining in multi-antenna radios. Modern radio systems often use multiple transmit antennas and appropriate signal coding to improve throughput and reliability of wireless links. Ideally, each amplifier should be rated at the maximum total (or sum) power that can be transmitted from all antennas. Due to cost and efficiency considerations, the power rating is usually designed to be lower (typically the maximum total power divided by the number of antennas), and this can lead to significant performance loss. This research explores a solution using Power Combining Networks (PCN) that enable the output of an amplifier to be switched/combined into any antenna, thus moving power from a weaker to a stronger antenna, resulting in improved link performance.The maximum output power of a MIMO (Multiple Input, Multiple Output) antenna transmitter is limited due to bio-safety considerations. MIMO transmission performs best when each power amplifier is rated at the full system transmit power. If the amplifiers are fractionally rated due to cost and efficiency considerations, the worst case loss with M antennas can reach log M dB in the presence of channel imbalances. A PCN can reduce this loss significantly, trading the loss of array gain for a better channel coupling efficiency. Use of a PCN alters the effective MIMO channel and therefore can also change the optimum transmit coding. The goal of this research is to study fast algorithms for PCN selection, appropriate signal coding schemes and the best realizable performance.