Research On Key Technologies Of Resource Management In Heterogeneous Cellular Networks

Recently, the surging traffic demands which are initiated by rapid growth of smart handsets in particular smartphones and tablets, have presented great challenges for wireless cellular networks. Furthermore, it has to be mentioned that the compelling developed wireless internet ap-plications, such as high definition video streaming and cloud compute, have been driving the great enhancement of network capacity and quali-ty of user experience. This gives a "huge shock" to the existed wireless communication systems. Especially, the cellular networks facing with the bottleneck of capacity enhancement because of the limited and ex-pensive sprctrum resources. Fueled by the urges of network capacity improvement and indoor coverage enhancement, HetNet has attracted significant attentions since it was proposed. HetNet is implemented by deploying LBSs which are underlayed the coverage of macrocell. LBSs have low signal transmit power, low cost and small coverage. Deploy LBS can highly improve the network coverage, mitigate coverage leak and enhance network capacity. The "everywhere and everytime" high speed access services can be provided to support the transmission of voice, image and media data. However, the HetNet is deployed diversely and densely. The resource management for macrocell network can’t be used in HetNet. The research on resource management in HetNet is nec-essary for its crucial role in network performance improvement.In this dissertation, dynamic BCA scheme with the consideration of base station load, eICIC and joint resource allocation network technolo-gy are studied. Correspondingly, several innovative schemes are studied and the performance of the proposed schemes is verified by theoretical analysis and computer simulation methods. The main contributions of this dissertation are concluded as the following:1) Considering the problem of load balancing in HetNet where the LBSs are underlaid deployed within the macro cell base station coverage. We design a dynamic BCA scheme for HetNet where each BS aims to balancing its load. The algorithm has low computation complexity and can achieve high SE and low transmission delay.Firstly, we model the traffic distribution as an inhomogeneous Poisson process and use the PPP, an uniform distribution statistical mod-el which is in the two dimensional plane, to model the BS placement. Then, the biasing coverage is defined based on the maximum biased SINR association policy. Moreover, we formulate our problem into a logarithmic utility objective function towards the throughput-optimal.Secondly, QPSO based cell association algorithm is developed. The algorithm includes LBS-MBS side which performances the suboptimal BCA solution and User-LBS side which performances the maximum bi-ased SINR based cell association.Finally, we compare the proposed dynamic BCA strategy in the metric of network average delay and cell spectral efficiency with the static BCA strategy.2) A SE enhancement scheme is proposed in HetNet by jointly con-sidering BCA and eICIC. The correlation of BCA factor and elCIC fac-tor with SE is investigated. The conclusions are useful for practical HetNet deployment to improve the total network SE.Firstly, we consider an OFDM (Orthogonal Frequency Division Multiplexing, OFDM) access based HetNet. As is modeled by Dirichlet tessellation, the MBS average coverage is obtained firstly. And then the biased cell radius of LBS is derived. The small cell coverage can be di-vided into centric and expanded area, and thus three parts are included by taking the MBS coverage into account. Based on this, the average number of macrocell users, small cell centric users, which are located at the unbi-ased area; and biased small cell users, which are located at the expanded area are derived hereinafter.Secondly, we focus on the definition of global network SE. Incor-porating with spectrum and power assignment, the network capacity can be obtained. However, the derived network capacity expression is in-tractable for the analysis of SE. Then, we introduce the transmission successful probability and propose a feasible method in defining net- work SE.Thirdly, we firstly prove that the macrocell and small cell centric area capacities are strictly decreasing of eICIC factor, and the throughput of small cell expanded area is strictly-increasing of eICIC factor. For the unavailable closed-form of optimal eICIC factor, we propose a bisection search based eICIC optimization algorithm then.Finally, we give numerical simulation results, which indicate our scheme can highly improve the network SE and obtain the approximate optimal eICIC factor by the proposed algorithm.3) We propose the joint resource allocation problem with QoS pro-visioning in downlink HetNet.Firstly, taking the different QoS demand of users’ services, we model the mixed integer programming problem towards capacity maximization.Secondly, the formulated problem is nonconvex with the integer constraint. We develop a two-stage solution. The integer subchannel as-signment and Lagrangian-based power allocation is designed.Finally, Simulation results show that the proposed online algorithm achieves a good tradeoff between network capacity and interference. Moreover, the average network efficiency is highly improved and the outage probability is decreased.