缓存使能毫米波大规模MIMO低时延无线传输理论研究

Millimeter wave (mmWave) massive multiple-input multiple-output (MIMO) technology has become a promising solution for the future mobile communication systems, since it can effectly utilize the space and spectrum resources. In order to satisfy the low latency requirements, the novel wireless transmission schemes with ultra-high rate and low latency are desired so as to balance the deploment cost, transmission latency and rsdio frequency (RF) chains overhead. In this project, base on cache-enabled ultra-dense network architecture, the wireless transmission theory of low latency for mmWave massive MIMO systems are investigated. As the existence of cache introdeces a new communication resources, we firstly consider content-centric based cache-asisted cloud edge coordinated transmission mechanism. Then, combining the cache-enabled ultra-dense networks and mmWave massive MIMO technology, the new spatial division multiplex technology in the case of the line of sight (LoS) MIMO channels and cloud edge coordinated transmission mechanism in the mmWave hybrid architecture are studied. Finally, aiming at the coordinated transmission among the edge nodes, we porpose the new distributed algorithm in the limited number of RF chains, which can decrease the transmission latency caused by the lots of information interaction among the edge nodes. This project can provide the fundamental theoretical of low lantency for the cache-enabled ultra-dense networks.

毫米波大规模MIMO技术能有效挖掘空间维度和频谱资源，成为新一代移动通信系统的关键技术。为满足未来移动通信的低时延需求，亟待探索超高速低时延的新型无线传输机制，以实现网络部署成本、传输时延以及射频链路开销的最佳平衡。本项目基于缓存使能超密集网络架构，拟开展毫米波大规模MIMO低时延无线传输理论研究。由于缓存的引入使通信系统获得了新的维度资源，本项目首先在缓存维度上，探析以内容为中心的缓存辅助云边传输机制。其次，融合毫米波大规模MIMO技术和缓存使能超密集网络，探索视距MIMO信道的空分复用和毫米波混合构架下的云边联合预编码设计。最后，针对边缘节点间的协作传输问题，提出新的分布式优化算法，考虑在射频通道数目受限条件下降低由边缘节点间信息交互而引起的传输时延。本项目的研究为缓存使能超密集网络的低时延传输奠定理论基础。

为应对移动互联网和“互联网+”的需求，未来移动通信无线传输理论和技术有待继续突破，低时延、高可靠和大带宽是未来移动通信的三大关键技术指标。将各种不同功能和特性的接入节点融合形成新的通信网络是提升上述指标的关键技术之一。目前，云接入、边缘计算以及直通通信等通信方式已深入研究。本项目研究了在云边端融合下的新型网络中的资源分配与调度、波束成型以及计算卸载策略等问题。针对典型场景下的通信模型，重点研究智能反射面辅助下的多小区协作收发波束成型设计，在高速传输的基础上实现了网络的低成本广覆盖。针对时延敏感性任务，在边缘计算网络中提出了流水线方式的计算卸载策略，有效降低了网络的计算与通信时延。针对异构网络中资源分配与调度，根据网络特性、信道状况以及信号格式，提出了联合资源分配与信道绑定方案，提升了多小区多节点下的系统容量界。研究成果为缓存使能的低时延通信系统提供了理论和技术支撑.