Research On Interference Coordination Based On Graph Theory For LTE HetNet

Long Term Evolution (LTE) is proposed in order to meet the rapidly increasing demandfor high speed data service and large system capacity in mobile communication system. Theinterference between cell-edge users in adjacent cells has become an extremely seriousproblem as the OFDMA employed in downlink access. Inter-cell interference (ICI) canseriously affect the overall system performance, in particularly, give rise to poor quality ofservice and low throughput for cell-edge users. It becomes a critical task of research toreduce ICI and improve the performance of cell-edge users. In addition, the development ofFemtocell is promoted by the requirements of larger coverage and higher data speed forindoor access. As Femtocells introduced, more cell-edges are added to cellular thereforecause more serious ICI in heterogeneous networks (HetNet). Interference management inHetNet has become the focus of studies.In this thesis we first outline the development and technical overview of LTE，then westate a survey about the ICI research, and analysis the ICI carefully in the background ofC-RAN access and HetNet. The downlink interference management consists of two parts inthis thesis, one is Inter-cell Interference Coordination (ICIC) and the other is resourceallocation. Graph theory is used to solving ICIC issues and UEs are considered as the vertexin the interference graph. We construct a simulation platform of the ICI in Macrocell system,and the simulation result denotes that the ICI scheme using graph theory has1dB advantageto the traditional system with reuse-1in SINR.Furthermore, the ICI in HetNet with Femtocells is studied in this thesis, and a systemlevel HetNet ICI simulation platform is designed. We modify the design model of theinterference graph against heterogeneous characteristic of HetNet, which improves the SINRperformance of both Macro and Femtocell users. Meanwhile, the Closed Subscriber Groupattribute is also researched for analyzing its effect to HetNet ICI. Finally, we compare theperformance of different algorithms at different system load.