【作者】 徐鹏；

【导师】 方旭明；

【作者基本信息】 西南交通大学 ， 通信与信息系统， 2011， 博士

【摘要】 随着电信产业与因特网的迅速发展,出现了种类繁多的无线通信网络系统如蜂窝通信系统、无线局域网、卫星通信系统、数字多媒体广播系统等,由其构成的混合网络共同为用户提供了泛在的无线环境,使用户可在任何时间、地点与任何个人进行通信。另一方面无线技术的多样性及其不断向前演进,造成了混合网络在结构和部署方式上具有复杂性与多样性的特点,如同代技术不同制式间(UMTS/WiMAX)、不同代技术间(2G/3G)、不同技术不同制式间(UMTS/WLAN)及同代技术不同规模间(Macro/Femtocell)等方式。如何充分利用各种无线通信网络间的互补特性,保证用户间通信的服务质量,实现真正意义上的混合网络自组织、自适应和无线资源管理技术的有机融合,是研究混合无线网络的关键所在。基于混合网络的复杂性,本文将混合网络分为异构网络与分层网络,采用博弈论、最优化理论等多种研究方法和手段,围绕混合无线网络资源管理中的接入控制、移动性管理、频谱分配及功率控制等问题进行深入的分析和研究,以达到提升混合网络资源效率、系统容量等系统性能的目的。首先对异构网络选择接入控制问题进行了研究,论文分析了现有异构网络选择算法和模型的优点与不足,提出了两种基于博弈论的新网络选择接入模型。区别于已有异构网络选择机制很少兼顾用户与网络双方,两种模型均综合考虑了用户与网络双方的状态及不同的需求目标,通过博弈进行相互选择并分别得到了博弈的均衡解。算法性能与数值分析结果表明两种模型在不牺牲用户与网络任何一方利益的条件下,分别在收益和满意度上取得了较高的系统性能增益,即在某种程度上使双方达到了一种双赢。随后研究了宏蜂窝/毫微微蜂窝分层网络中的切换判决问题,在分析了已有切换算法在分层网络中由于毫微微蜂窝极低发射功率的特性而不能进行有效地判决之后,提出了两种适用于分层网络的新切换算法。在联合信号强度与功率损耗的切换算法中,采用接收信号强度与功率损耗进行联合切换判决,可以有效地提高毫微微蜂窝的利用率和解决宏蜂窝与毫微微蜂窝基站功率的差异问题,同时对用户起到一定的节能作用,性能与数值分析结果表明该算法对比于传统算法增加了切换次数,对比于其他分层切换算法减少了一些不必要的切换次数。在基于信干噪比和用户状态的切换算法中,通过信干噪比避免基站间的功率不对称性,采用用户状态包括移动速度、业务服务质量需求来避免仅依据信干噪比判决引起的频繁切换,仿真分析结果表明该算法提高了毫微微蜂窝的利用率,降低了造成频繁切换的可能性,从而提升了系统性能。接着研究了毫微微蜂窝部署于宏蜂窝中时具有同样授权频段的基站间该如何分配频谱的问题,在分析了传统宏蜂窝采用的集中式频谱规划/分配机制在分层网络中由于毫微微蜂窝有线回程的特点不再适用后,提出了分层网络中基于博弈论的两种自适应频谱分配机制。两种机制分别结合静态矩阵博弈理论与动态斯坦伯格博弈理论,将宏蜂窝和毫微微蜂窝基站分别作为博弈双方,共同的频谱资源作为基站间博弈竞争的对象,通过求解博弈均衡以最小化跨层、同层干扰,同时给出了基于矩阵博弈的分配机制中当均衡不存在时的次优解和基于斯坦伯格博弈的分配机制中均衡的定义及存在性证明。分析结果表明两种自适应频谱分配机制均实现了频谱的混合分配方式,即在频谱资源充足时进行正交分配,在资源匮乏时进行复用／共道分配,以最大程度地提高频谱利用率。性能与数值分析结果表明两种自适应机制在吞吐率、阻塞率与频谱效率上都取得了比较明显的性能改善。最后对分层网络中的毫微微蜂窝自配置的相关问题进行了研究,基于毫微微蜂窝最终将由用户进行自主安装和配置且大多数用户并不具有相关专业技术背景的特点,结合最优化理论提出了一种毫微微蜂窝基站功率的自配置机制。该机制在分析部署新毫微微蜂窝基站时可能受到的各种类型的干扰和可能引起对已有用户和网络的干扰类型的基础上,根据分析得到的干扰限定条件设计了一个自配置功率的最优化问题,即在保证已有用户正常通信的条件下如何最大化系统容量。分析结果表明,通过最优化理论中的拉格朗日乘数法可以求出该机制中发射功率设置的最优解。仿真比较结果表明自配置机制实现了毫微微蜂窝基站依据周围无线环境对发射功率的自适应调整,有效地降低了毫微微蜂窝的发射功率,减少了添加毫微微蜂窝可能引起的各种干扰,从而提升了系统性能。更多还原

【Abstract】 With the rapid development of telecom industry and Internet, various kinds of wireless communication networks have appeared such as radio cell communication systems, wireless local networks, satellite communication system, and digital multimedia broadcasting systems, etc. Hybrid network composed of these systems provides ubiquitous radio environment for users, which makes people commnunicate with anyone at any time and any place. On the other hand the diversity and constant evolution of wireless technology led to the complexity and multiplicity of hybrid network on construction and deployment, for example, different generation technologies (2G/3G), same generation with different standards (UMTS/WiMAX), different technology and different standard (UMTS/WLAN), and same generation with different scale (Macro/Femtocell). The key for reasearching hybrid network includes that how to sufficiently utilize the complementary property of various wirless networks for guaranteeing users’QoS, how to truly implement the self-organization and adaptation of hybrid network, and how to integrate radio resource management of different networks. The hybrid network is divided into heterogenous and hierarchical networks due to its complexity, research issues of hybrid wireless network resource management such as admission control, mobility management, spectrum allocation and power control are analyzed and investigated thoroughly by using game theory and optimization theory in this dissertation, in order to improve the system performance like resource efficiency, capacity and so on.Firstly, the issue on heterogenous network selection is investegated. The advantages and disadvantages of existing related models and algorithms are analyzed, and then two new models based on game theory are proposed. Different from existing schemes which hardly give attention to both users and networks, the status and different requirements of both users and networks are taken into consideration in the two new models, and the equilibrium points of two games can be achieved through mutual selection. Simulation results show the two models can separately achieve higher gain of system performance on revenue and satisfaction degree under the condition of without sacrificing users or networks profits. That is to say, users and networks attain a win-win situation.Secondly the issue on handoff decision in Macro/Femtocell hierarchical networks is researched. Based on the lack of existing handoff algorithms, due to the property of femtocells which has very low transmit power, two new handoff algorithms in hierarchical network are proposed. The handoff decision is made by adopting received signal strength (RSS) combined with power loss in the first algorithm, which can improve the utilization of femtocells and solve the discrepancy on transmit power of macro and femtocell base stations, and make some power-saving. Simulation results show the algorithm increases the number of handoff compared to conventional algorithms, and ruduces the number of unnecessary handoff compared to other hierarchical handoff algorithm. In the signal to interference plus noise ratio (SINR) based algorithm, the SINR used to avoid power asymmetry of base staions and users’state including velocity, and the QoS used for decreasing frequent handoff are combined. Simulation results show the second algorithm promotes the utilization ratio of femtocells, diminishes the possibility of over handoff, therefoer improves the system performance.Thirdly the dissertation researches the issue how to allocate spectrum among base stations in hierarchical network when femtocells are deployed in macrocells and they share the same licensed frequency band. The centralized spectrum planning and allocation schemes do not work efficiently in hierarchical network owing to the wired backhaul characteristic of femtocells, thus two game theory based adaptive schemes of spectrum allocation are proposed. Two schemes integrate the theory of matrix game and Stackelberg game, respectively, both of which make the base stations of macrocell and femtocell as the two sides of players, and make the same spectrum as the game object, to minimize the co-layer and cross-layer interferences by solving the equilibrium of the game. In addition, the suboptimal solution is given when the equilibrium of matrix game does not exist. The definition and existence proof of Stackelberg game equilibrium are also given. Analysis results represent both of the proposed schemes allocate spectrum in a hybrid way, namely allocating orthogonally when resource is enough and in a reuse or co-channel manner when resource is scarce. Simulation results show the two adaptive schemes attaine the obvious performance improvement in throughput, blocking probability, and spectrum efficiency.At last the attention is paid to the issue on self-configuration of femtocells. As the femtocells are finally deployed independently by users without related professional and technical background, a femtocell self-configruation scheme base on optimization theory is proposed. Based on analyzing all kinds of received interference and types of introduced interference on existing users and networks, an optimal problem of power self-configuration is designed according to interference qualification, which refers to maximize system capacity on the condition eusuring normal communication of existing users. Analysis results present the transmit power optimal solutioan can be obtained with Lagrange multiplier of optimization theory. Simulation results demonstrate the scheme realizes transmit power adaptation of femtocells in accordance with radio environment, effectively cuts down transmit power, reduces various introduced interference, and enhances system performance.更多还原