Precision Measurement and Modeling of Dynamic Millimeter-wave Wireless Propagation Channels

5G通信的一个决定性特征是使用毫米波范围内的频段。充足的可用带宽有可能实现更高的数据速率，从而实现大量新应用，从改进的视频流到虚拟现实再到工业监控。然而，要正确评估这种毫米波系统的潜力和局限性，首先需要了解传播信道，即信号从发射机传播到接收机的方式。由于衍射和散射等基本传播效应在更高的频率下显著不同，因此可以预期整体传播通道与传统蜂窝频率下已探明的通道不同。拟议的项目将提供毫米波传播通道的详细、基于测量的描述，特别强调由环境中移动物体（汽车、人类、机械）产生的时间变化。从这样的理解中，有可能获得如何设计更可靠、更有效的毫米波通信系统的见解。由于毫米波通信的重要性，已经存在了许多针对毫米波信道的测量方法，但它们都表现出严重的局限性。特别是，没有测量可以同时(i)提供高分辨率的方向信息，（ii）是动态的，即显示移动设备或散射物体的影响，以及（iii）提供统计上显著数量的测量点，这些测量点可以形成随机通道模型或机器学习训练的可靠基础。由于缺乏测量结果，5G设备和系统开发中使用的许多假设都是猜测，本项目旨在证明或推翻这些假设。为了实现这一目标，该项目将使用南加州大学最近开发的一种新型信道测深仪，并通过先进的信号处理技术扩展其功能。该信道测深仪基于快速波束开关原理，可实现高等效各向同性辐射功率（EIRP），并在短时间（10ms）内捕获完整的定向信道特性。使用该探测仪，项目将执行和评估广泛的测量活动，其中一些将集中于动态影响和非平稳性，而另一些将利用测量和评估大量数据的能力。与被广泛引用的现有测量方法相比，新测量方法的速度要快100万倍，测量位置要多三个数量级。该项目的另一个重要成果将是开发新的渠道模型，这些模型可以反映理论分析和系统设计的所有相关渠道属性。通过关注结果的空间一致性，并分析多径分量的数量和幅度分布，可以更好地进行部署规划，并对系统性能产生影响，例如预测各种波束形成器架构。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

One of the defining features of 5G communications is the use of frequency bands in the mm-wave range. The ample available bandwidth has the potential to enable dramatically higher data rates, thus enabling a plethora of new applications, ranging from improved video streaming to virtual reality to industrial monitoring and control. However, to properly assess the potential and limitations of such mm-wave systems, it is required to first understand the propagation channel, i.e., the way in which signals propagate from the transmitter to the receiver. Since the fundamental propagation effects such as diffraction and scattering are significantly different at higher frequencies, the overall propagation channel can be expected to be different from the well-explored channels at traditional cellular frequencies. The proposed project will provide a detailed, measurement-based description of mm-wave propagation channels, with special emphasis on the time variations that are created by moving objects (cars, humans, machinery) in the environment. From such understanding, it is possible to obtain insights in how to design more reliable, and more efficient, mm-wave communications systems. Due to the great importance of mm-wave communication, a number of measurements do exist for mm-wave channels, but they show serious restrictions. In particular, no measurements are available that simultaneously (i) provide directional information with high resolution, (ii) are dynamic, i.e., show the impact of moving devices or scattering objects, and (iii) provide a statistically significant number of measurement points that could form a reliable basis of stochastic channel models, or training for machine learning. Because of a lack of measurement results, many assumptions that are used in the development of 5G devices and systems are conjectures, which this project aims to prove or disprove. To achieve this, this project will use a novel channel sounder recently developed at University of Southern California and extend its capabilities through advanced signal processing techniques. This channel sounder is based on the principle of fast beamswitching, which enables high equivalent isotropically radiated power (EIRP) and capturing complete directional channel characteristics within a short time (10ms). Using this sounder, the project will perform and evaluate extensive measurement campaigns, some of which will concentrate on dynamic effects and nonstationarities, while others will exploit the capability for measuring and evaluating massive amount of data. Compared to widely cited existing measurements, the new measurements can be done one million times faster, and three orders of magnitude more measurement locations. Another important result of the project will be the development of new channel models that can reflect all of the relevant channel properties for theoretical analysis as well as system design. By paying attention to the spatial consistency of the results, and analyzing the number and amplitude distribution of the multipath components, better deployment planning, and impact on system performance such as prediction of various beamformer architectures will be enabled.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.