Research On Interference Coordination In LTE-A Heterogeneous Networks

With the rapid development of mobile Internet, mobile users put forward a higher demand for the data rates. Applications such as online games, video calls, high-speed downloads have led to the increasing demand for higher data rates and lower latency of the cellular networks. On the other hand, it is estimated that2/3of calls and over90%of data services occur indoors. However, due to the building penetration loss, more power from the cellular base station will be needed to provide indoor coverage. Therefore,3GPP (The3rd Generation Partnership Project) has introduced low power nodes in LTE (Long Term Evolution) Release10to make users get closer to base stations, thus improve the user throughput. The architecture of deploying low power nodes on the basis of traditional macrocell network is called Heterogeneous Networks (HetNets).In order to maximize the use of spectrum resources,3GPP tends to share the same frequency band between the macrocell and small cells. The co-layer interference among macrocells and that among small cells, as well as the cross-layer interference between macrocells and small cells have become an important factor in restricting the performance of HetNets. Compared to the traditional macro base stations, low power nodes are usually deployed randomly with the closed access mode and range expansion technology, which makes the interference problem of HetNets more complicated. Therefore, the Inter-Cell Interference Coordination (ICIC) technology is one of the most important technologies in LTE-A HetNets. In this dissertation, taking the typical HetNets scenarios Macro-Femto and Macro-Pico networks as an example, we study the ICIC technologies in LTE-A HetNets. The main contributions and innovations are listed as follows:1) This dissertation studies the downlink interference problems of the Macro-Femto network, and proposes an interference graph based frequency-domain ICIC scheme.In the construction of the interference graph, the vertexes are all femto base stations and the users which are in the interference regions of femtocells, the edge represents the interference relationship between two vertexes, and the weight represents the degree of the interference. The weight of edges is characterized by the Regional Average Channel State (RACS) metric. Then we model the optimal clustering problem as the MAX K-CUT problem in the graph theory, and propose a heuristic algorithm named Hybrid Clustering based on Interference Graph (HCIG). The HCIG algorithm could reduce all the three types of interference in Macro-Femto networks and improve the network throughput.2) This dissertation studies the co-layer interference problem among femtocells, and proposes an ICIC scheme based on the cooperation between femtocells.In the3GPP LTE standards, there is an X2interface between neighbor femtocells, which makes the cooperation between femtocells become possible. By enabling a femtocell to relay the data of its neighbor femtocell, we propose a virtual handover scheme to cancel the interference between femtocells. The main idea of virtual handover is to separate the user plane and control plane. When the virtual handover occurs, the control messages are transferred by the original serving femtocell, and the data messages are transferred by the aided neighbor femtocell. The virtual handover is transparent to the user and the Mobility Management Entity (MME), and it only needs the control message exchange between femtocells. Furthermore, the message exchange is through the wired X2link, which makes the virtual handover scheme have a low overhead and a high probability of success.3) This dissertation studies the interference from the macrocell to picocell users, discusses the ratio of Almost Blank Subframe (ABS) in the ABS-based time-domain ICIC scheme.When Range Expansion (RE) is used in the Macro-Pico network, the users in picocell expanded regions will suffer high interference from the macrocell. Generally, ABS-based time-domain ICIC scheme is utilized to cancel this type of interference. We use the Poisson Point Process (PPP) model to analyze the performance of the Macro-Pico network with RE, and get the distribution of distance between users and base stations, SIR, average user throughput. Then we get the reasonable range of ABS ratio when the range expansion bias is given. The work is helpful to the settings of range expansion bias and ABS ratio in the practical system.4) This dissertation studies the impact of range expansion on the user outage probability, proposes a load-aware PPP model.Observed that the load of base stations is not taken into account in the traditional PPP model, which is simply assumed that all base stations are fully loaded, we take the effects of load and QoS into the consideration, and propose a more practical load-aware PPP model. Based on the proposed model, we analyze the impact of range expansion on the user outage probability, and get the outage probability expression when the range expansion bias is given.5) This dissertation studies the interference problem of the Macro-Pico network, proposes a load adaptive ICIC scheme.In order to make picocells provide more communication resources to the hotspot users, we propose a hotspot-based bounded range expansion scheme to expand the coverage of picocells. Based on the bounded range expansion scheme, we propose a Load-adaptive Dynamic Range Control (LDRC) algorithm to achieve the load balancing between the macrocell and picocell, and give a hybrid approach of time and frequency domain ICIC to reduce the interference from macrocell to users in picocell expanded regions. The LDRC and proposed ICIC scheme could achieve a good trade-off between QoS guarantee and system capacity.