面向数据分发的蜂窝直通通信技术研究

发布时间：2018-09-03 18:24

【摘要】：随着信息技术与无线网络技术的不断发展,无线多媒体业务已经呈现出爆炸式增长趋势。在现有网络条件下,如何更有效利用有限网络资源实现高速率数据分发己成为无线通信研究领域内最重要的课题之一。为满足不断增长的新业务及新需求,端到端直通通信(Device-to-Device, D2D)技术做为一种高效的信息传输方案,被认为是移动互联网领域最有前途的大规模数据分发技术之一。D2D技术的核心思想是利用D2D通信速率高、频谱效率高、功耗低等优势完成高效的点对点或点对多点传输,不再通过基站转发。国内外众多知名高校、科研机构和公司都积极投入对D2D的研究,并取得了些重要的研究成果。经证明,D2D能显著改善数据分发系统性能,如提升网络吞吐量、提高频谱效率、降低网络负载等。而当D2D引入到传统的蜂窝网络系统后形成D2D辅助蜂窝网络,网络环境变得更加复杂。因此,需要对面向数据分发的D2D通信网络进行优化,以减少用户之间相互干扰,满足用户QoS需求,对进一步提升D2D辅助蜂窝网络数据分发的性能并促进D2D的大规模应用具有重要的指导意义。本论文首先研究海量D2D用户共存场景下以提高D2D网络可靠性为目标的资源分配算法；其次研究D2D辅助蜂窝网络中在保障用户QoS需求和收益条件下,以提高混合网络容量为目标的资源分配算法,包括正交频谱划分和全局频率复用两种场景；最后考虑多播协作用户间的紧密性,研究D2D协作多播数据分发算法。本论文的主要贡献包括以下几个方面：第一,在多个D2D用户在有限区域内共存的场景下,首先利用随机几何理论对D2D网络建模,将D2D用户建模为在有限平面上用户数目服从时间维度上的泊松分布,在二维平面位置上服从均匀分布的PPP过程。在同时考虑路径损耗和阴影衰落的条件下,推导了多个D2D用户共信道部署情况下的中断概率表达式,并分析了影响中断概率的关键参数。其次,针对D2D用户在有限区域内密集部署的场景,提出了提高D2D传输可靠性的频谱分配与功率控制算法。最后仿真验证了所提出的资源分配算法在传输可靠性方面有明显优势。第二,在D2D辅助蜂窝网络正交频率划分场景下,有若干蜂窝用户与D2D用户在有限区域内使用相互正交资源,研究了基于用户差异化QoS保障的资源分配算法。首先,针对优先级较高的宏蜂窝用户提出满足其最低速率需求的最小信道分配策略。其次,针对D2D用户的传输可靠性要求,理论分析了共信道部署的最大用户密度,基于该理论分析提出了基于用户分簇的启发式资源管理算法。最后,以最大化网络容量为目标按D2D用户簇进行信道分配。仿真表明,所提出的资源分配算法不仅能保障蜂窝用户QoS需求,同时也能满足D2D用户中断概率的限制。第三,在D2D辅助蜂窝网络中全局频率复用场景下,若干蜂窝用户与一个D2D用户复用全部资源。研究了全局频谱复用下基于网络收益保障的频谱与功率联合分配算法。首先,对D2D用户在每个子信道上的支付函数以及蜂窝网络的成本函数建模,建立网络收益模型。其次,建立了以最大化混合网络吞吐量为目标,以满足网络收益为条件的子信道分配和功率分配的联合优化问题模型。最后,利用凸优化理论提出了基于对偶分解的最优化算法和两步走的次优化算法。仿真结果表明,所提最优化算法与传统资源分配算法相比,在网络容量方面有明显提升。此外,次优化算法在极大降低运算复杂度的同时,网络容量接近全局最优解。第四,为提高蜂窝网络中数据分发效率,研究了基于D2D协作多播的数据分发策略,该策略分两个阶段。第一阶段,基站首先根据多播业务速率要求进行多播。第二阶段,通过簇内D2D协作多播向未正确接收的用户进行数据重传。首先,为提高协作多播的可靠性,设计统一的用户紧密度评价指标,并基于该指标构建表征用户紧密度的邻居图。其次,提出一种以最大化网络吞吐量为目标的分簇策略。最后,在两个阶段都没有成功接收数据的用户直接向基站请求数据。仿真结果表明,该数据分发策略不仅提高了网络容量,并且极大降低了基站负载。
[Abstract]:With the continuous development of information technology and wireless network technology, wireless multimedia services have shown an explosive growth trend. Under the existing network conditions, how to make more effective use of limited network resources to achieve high-speed data distribution has become one of the most important topics in the field of wireless communication research. As an efficient information transmission scheme, end-to-end direct communication (D2D) technology is considered as one of the most promising large-scale data distribution technologies in the field of mobile Internet. The core idea of D2D technology is to make use of the advantages of high speed, high spectrum efficiency and low power consumption of D2D to complete efficient point-to-point or point-to-point communication. Point-to-multipoint transmission is no longer transmitted through base stations. Many well-known universities, research institutions and companies at home and abroad have actively engaged in the research of D2D, and have achieved some important research results.
It has been proved that D2D can significantly improve the performance of data distribution system, such as increasing network throughput, improving spectrum efficiency, reducing network load, etc. However, when D2D is introduced into traditional cellular network system, the network environment becomes more complex. Therefore, it is necessary to optimize the data distribution-oriented D2D communication network to reduce network load. It is of great significance to further improve the performance of D2D-assisted cellular network data distribution and promote the large-scale application of D2D.
Firstly, this paper studies the resource allocation algorithm aiming at improving the reliability of D2D network in the scenario of massive D2D users coexistence; secondly, it studies the resource allocation algorithm aiming at improving the capacity of hybrid network under the condition of guaranteeing the QoS demand and revenue of users in D2D-assisted cellular network, including orthogonal spectrum partitioning and global frequency reuse. Finally, considering the tightness of multicast cooperative users, the D2D cooperative multicast data distribution algorithm is studied.
Firstly, in the scenario where multiple D2D users coexist in a finite area, the D2D network is modeled by using stochastic geometry theory. The number of D2D users is modeled as a Poisson distribution on a finite plane and a PPP process with uniform distribution on a two-dimensional plane. Under the condition, the interruption probability expression is deduced and the key parameters affecting the interruption probability are analyzed. Secondly, a spectrum allocation and power control algorithm is proposed to improve the reliability of D2D transmission in the scenario of densely deployed D2D users in a limited area. The source allocation algorithm has obvious advantages in transmission reliability.
Secondly, in the scenario of D2D-assisted orthogonal frequency division in cellular networks, some cellular users and D2D users use orthogonal resources in a limited area, and a resource allocation algorithm based on user-differentiated QoS guarantee is studied. Firstly, a minimum channel allocation policy is proposed to satisfy the minimum rate requirement of high-priority macro cellular users. Secondly, aiming at the transmission reliability requirement of D2D users, the maximum user density of common channel deployment is theoretically analyzed, and a heuristic resource management algorithm based on user clustering is proposed. Finally, the channel allocation is carried out according to D2D user cluster with the goal of maximizing network capacity. It not only guarantees the QoS demand of cellular users, but also meets the D2D user outage probability limit.
Thirdly, in the global frequency reuse scenario in D2D-assisted cellular networks, several cellular users and a D2D user reuse all the resources. A joint spectrum and power allocation algorithm based on network revenue guarantee is studied in this paper. Firstly, the payment function of D2D users on each subchannel and the cost function of cellular networks are constructed. Secondly, a joint optimization model of subchannel allocation and power allocation is established to maximize the throughput of hybrid networks and satisfy the network benefits. Finally, an optimal algorithm based on dual decomposition and a two-step suboptimal algorithm are proposed by using convex optimization theory. Compared with the traditional resource allocation algorithm, the proposed optimization algorithm can significantly improve the network capacity. In addition, the sub-optimization algorithm can greatly reduce the computational complexity while the network capacity is close to the global optimal solution.
Fourthly, in order to improve the efficiency of data distribution in cellular networks, a data distribution strategy based on D2D cooperative multicast is studied. The strategy is divided into two stages. In the first stage, the base station performs multicast according to the multicast traffic rate requirement. Secondly, a clustering strategy is proposed to maximize network throughput. Finally, users who have not received data successfully in both stages request data directly from the base station. Obviously, the data distribution strategy not only improves the network capacity, but also greatly reduces the load of the base station.
【学位授予单位】：北京邮电大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN929.5

【参考文献】