面向超密集组网的干扰压缩理论和方法研究

Ultra dense network(UDN) is widely considered as a key component of 5G mobile system, which deploy low power transmission points with extremely high density in hotspot and can improve ~100 times of the system capacity. But system capacity of UDN is not necessarily proportional to cell density as densification, the interference is the main bottle neck of the performance improvement. With denser cells, the interference is much more, more varying and more difficult to cooperate among adjacent cells as to lack a unified center node. We initiate the compressive domain interference cancellation for UDN. .Compressive domain interference cancellation combines the interference suppression and compressive sensing. It is able to efficiently filter out the interference from the compressive measurements in a manner that preserves the ability to recover the signal of interest. The construction operates completely in the compressive domain and has computational complexity that is polynomial in the number of measurements. Compressive domain interference cancellation assures UDN can achieve better performance gains through effective interference filtering. It is innovative for the academic research and valuable to 5G mobile system.

作为5G关键技术之一，超密集组网通过更加密集化的无线网络设备部署，使热点区域实现百倍量级的系统容量提升，但复杂、多变的干扰是实现其性能优势的主要瓶颈。针对超密集组网干扰源多、干扰复杂多变、小区间难以协作的特点，本课题提出了干扰压缩技术。干扰压缩将干扰抑制和压缩感知技术相结合，将混杂着大量干扰的原始信号，通过压缩感知采样将M维降为N维，其中M>>N，对降维后的目标信号进行干扰消除或干扰抑制处理，然后恢复重构出理想的目标信号。.干扰压缩后的信号维数远小于压缩前的原始信号，不仅降低了干扰抑制的复杂度，更易求得干扰抑制的最优解，能显著提升超密集组网的干扰抑制效果。最后通过合理的信号重构，将抑制了干扰的目标信号完整、无畸变地重现，保证了干扰压缩后信号传输质量。干扰压缩将实现超密集组网在干扰个数增加时性能不会明显降级，解决了超密集组网的瓶颈问题，具有良好的应用前景。

本课题将压缩感知与干扰抑制技术结合，以改善超密集组网的干扰抑制效果、减小抑制算法的运算复杂度、提升系统能效为目标，解决了超密集组网的瓶颈问题，显著提升系统容量，降低基站发射功率和功放制造成本。针对超密集组网的应用场景，对无线信道稀疏性进行分析，实现超密集组网场景下信号稀疏性的判定和一致性表达；针对小区干扰和宏-微场景，提出了基于最小均方误差的干扰抑制算法和基于正交投影的干扰抑制算法；针对微基站密集部署导致的干扰问题，提出了基于压缩感知的超密集网络干扰对齐和干扰抑制技术、基于超密集网络的分层预编码技术、基于改进DBSCAN聚类算法的超密集网络干扰抑制方案；提出适于干扰压缩应用场景的信号重构和恢复技术。验证了基于压缩感知框架的干扰抑制算法可以有效消除超密集网络干扰，提升了系统容量。在本项目执行期间，项目组在国际无线通信与信号处理权威期刊IEEE汇刊（WCL、IOT、TVT）以及电子信息学报、通信学报上发表论文21篇，在本领域重要的国际会议（IEEE GC、ICC、Mobicom、Infocom等）上发表论文8篇，其中包括ESI热点论文1篇、ESI高被引论文1篇，取得了较好的国际影响力。搭建超密集组网原型验证平台1套，申请国家发明专利21项。培养硕士研究生23名。