用于通信的微波线性模拟计算机（MiLACs）：机遇与挑战
Microwave Linear Analog Computers (MiLACs) for Communications: Opportunities and Challenges
浏览论文内容
中文总结 AI 辅助
研究未来无线系统中传统数字MIMO架构难扩展的问题，提出用微波线性模拟计算机MiLACs将部分处理卸载到模拟域的方法，阐述其可减少射频链数量等优势，并讨论了相关挑战及未来研究方向。
中文摘要 AI 辅助
未来无线系统需要更大天线阵列和更繁重信号处理，传统数字多输入多输出（MIMO）架构难以扩展。本文表明一种可能的解决方案是将部分处理从数字域卸载到模拟域，可通过线性微波网络（即微波线性模拟计算机MiLACs）直接利用射频通信信号计算。MiLACs能通过波传播立即执行有用矩阵运算，虽线性但输出信号可非线性依赖网络可调参数，可实现矩阵求逆等运算。接着回顾了MiLAC辅助MIMO架构能减少射频链数量等。最后讨论了与MiLAC相关的主要挑战及未来研究的有前景方向。
英文摘要
Future wireless systems will require ever larger antenna arrays and heavier signal processing, making conventional digital multiple-input multiple-output (MIMO) architectures difficult to scale. In this paper, we show that a possible solution is to offload part of the processing from the digital to the analog domain. This can be done through linear microwave networks designed to compute directly using the communication signals at radio frequency (RF). These networks, denoted as microwave linear analog computers (MiLACs), can perform useful matrix operations instantly through wave propagation. Remarkably, although MiLACs are linear, the output signals can depend nonlinearly on the tunable parameters of the network, enabling the computation of operations beyond simple linear transforms. In particular, MiLACs can realize matrix inversion and pseudo-inversion with complexity scaling quadratically with matrix size, rather than cubically, which is essential in zero-forcing beamforming. We then review how MiLAC-aided MIMO architectures can reduce the number of RF chains, relax the resolution requirements on digital-to-analog converters (DACs) and analog-to-digital converters (ADCs), and decrease the beamforming complexity. We finally discuss the main challenges related to MiLAC and promising directions for future research.
