Research Equipment for Intelligent Signal Processing on Millimeter Wave MIMO Communication in 5G Networks

由于第五代（5G）无线网络对更高的传输速率、链路可靠性、无处不在的互联网接入等更好的服务需求不断增长，在毫米波频率（30GHz-300GHz）下工作的多输入多输出（MIMO）系统最近受到了广泛关注。然而，MIMO（特别是当天线数量增加时）受到所有链路需要大量信道状态信息的限制，并且在上行链路中发射导频信号进行信道估计成为一项消耗资源且具有挑战性的任务。此外，毫米波频率下的稀疏信道在信号处理和通信领域也受到了独特的关注。因此，为了在即将到来的5G网络中应用毫米波支持的MIMO或大规模MIMO系统，必须开发智能信号处理技术。我目前的发现项目专注于为多用户大规模MIMO系统开发鲁棒和节能的信号处理技术。本研究的主要目标是在实际信道条件下，最大限度地提高大规模MIMO网络的传输可靠性、频谱利用率和能源效率。然而，由于缺乏真实的通道测量和实时测试平台，在发现项目下开发的算法在很大程度上是学术性的。此外，非线性功率放大器、实际天线和馈线引入的畸变也无法准确建模。

Motivated by the increasing demands for better services in the fifth-generation (5G) wireless networks such as higher transmission rate, link reliability, ubiquitous access to internet etc, the multi-input multi-output (MIMO) system operating at millimeter wave frequencies (30GHz-300GHz) has recently received a great deal of attention. However, MIMO (especially when the number of antennas increases) has been constrained by the need of extensive channel state information for all links and transmitting pilot signals in the uplink for channel estimation becomes a resource-consuming and challenging task. Moreover, the sparse channel under millimeter wave frequencies presents a unique concern in the field of signal processing and communication. As such, it is imperative to develop intelligent signal processing techniques for millimeter wave enabled MIMO or massive MIMO systems to be applied in upcoming 5G networks. My current Discover project is focused on the development of robust and energy efficient signal processing techniques for multi-user massive MIMO systems. The main objective of this research is to maximize the transmission reliability, spectral usage and energy efficiency of massive MIMO networks under realistic channel conditions. However, the algorithms developed under the Discovery project are largely academic due to the lack of real channel measurements and real-time testbed. Also, the non-linear power amplifiers, practical antennas, and distortions introduced by the feed lines could not be accurately modeled.