【作者】 魏垚；

【导师】 王文博；

【作者基本信息】 北京邮电大学 ， 通信与信息系统， 2013， 博士

【摘要】 无线网络自组织(Self-Organizing Network, SON)技术被认为是未来管理及维护复杂网络,有效支撑异构多制式、提高网络整体性能、大幅度降低运营成本的最有效途径,因此近年来受到了学术界和工业界的广泛关注。物理小区标识(Physical Cell Identity, PCI)和邻区关系列表(Neighbor Cell List, NCL)的自组织技术是SON的重要用例,它们实现了移动用户在小区接入过程中能够准确识别服务小区和非服务小区,在邻区切换过程中能够识别相邻小区和非相邻小区。小区识别技术随着蜂窝移动通信系统的演进而不断变化：从第二代全球移动通信系统GSM的频率复用技术,到第三代TD-SCDMA系统的扰码与频率规划结合、以及WCDMA系统的扰码分组配置技术,再到第四代系统应用物理小区标识区分小区,实现相邻小区“不冲突”和“不混淆”。随着蜂窝移动通信系统的演进,不同系统的小区识别技术各具鲜明的特征,但其通过复用来解决资源缺乏问题的核心思想一直未变。本论文以无线网络自组织作为研究背景,对物理小区标识和邻区关系列表的自配置与自优化技术进行了深入而全面的研究。文章结合前几代蜂窝移动通信系统的小区识别策略,提出了PCI和NCL的优化配置算法,并对比现存的技术方案,对物理资源特性,资源分配,组网性能,用户体验等方面进行了深入而全面的研究。论文的主要工作和创新之处在于：1.物理小区标识码特性研究首先,论文对物理小区标识PCI应用于小区搜索和同步过程中的码特性进行了研究,包括物理小区标识与同步序列的映射关系、同步序列码的生成及其独有的码特性等。论文重点针对主同步序列和辅同步序列的时频偏移相关性进行研究,分析了主辅同步序列的抗时间偏移和频率偏移的特性,研究发现辅同步序列在时域上受扰码的影响,存在多组高互相关的序列。论文分析了高互相关的辅同步序列的分布特征,并根据其分布特点进行分组规划,提出分组分配PCI的配置机制。2.单层同构场景下物理小区标识自组织技术研究本论文就长期演进增强(Long Term Evolution-Advanced, LTE-A)系统单层同构场景下的物理小区标识PCI自组织进行了研究。考虑到已有的分配算法仅满足相邻小区ID“不冲突”和“不混淆”的基本配置要求,并没有追求复用距离尽量远,复用干扰尽量小的最优化目标,因此论文采用了图论中最小生成树算法,把基站的PCI配置问题转化为图论问题,利用贪婪算法思想搜索出网络最大干扰,并使用正交PCI资源进行规避；同样采用贪婪思想搜索出在PCI资源复用时的当前最优解,以最小化网络中的复用干扰。论文还提出了一种基于复用频率的PCI自配置算法,算法通过网络邻接矩阵求解基站的配置顺序,再按照顺序配置复用频率最低的PCI,使复用干扰平均化。论文详细地分析了算法复杂度和性能,仿真验证结果表明：所提算法与已有的自配置算法相比具有更低的算法复杂度和更高的用户载干比性能。3.分层异构场景下物理小区标识自组织技术研究根据分层异构网络节点的部署特点,论文提出了适用于多类型节点分层密集部署网络环境的PCI着色置换法和PCI分类配置法。PCI着色置换法通过局部贪婪搜索算法思想对存在冲突的相邻节点ID进行交换或替换,逐步降低网络干扰；PCI分类配置法则采用IP地址分配的思想,将PCI资源分组并标号,网络节点可以根据邻区ID快速寻址定位自身的PCI。仿真验证表明所提配置算法都能保证宏基站之间,异构节点之间以及宏基站与异构节点间的PCI“不冲突”和“不混淆”,而且能有效降低网络复用干扰。4.邻区关系列表自组织技术研究论文对邻区关系列表的自组织技术进行了深入的研究。论文给出了基于信号传播预测的邻区关系列表自配置算法,以及针对由于小区覆盖动态变化而引起的邻区漏配,邻区冗余的自优化算法。算法采集小区边缘移动终端上报的测量信息和切换信息,通过预设的门限和邻区的生存时间进行漏配邻区删除和冗余邻区添加的判决,实现邻区关系的动态变化。论文通过搭建动态系统级仿真平台,对自动邻区关系算法进行仿真验证。仿真结果表明在经过邻区关系列表优化配置后,用户无线链路失败率均有明显改善,而用户切换成功率具有8%的性能增益。更多还原

【Abstract】 Self-Organizing Network (SON) is considered to be the most effective way to manage and maintain the future networks, to support the heterogeneous network, to improve the performance of the network and to reduce Operating Expense (OPEX) significantly, thus it has been widely concerned in academia and industry in recent years. Physical Cell Identity (PCI) and Neighbor Cell List (NCL) are important cases in SON, which enables User Equipment (UE) to identify serving cell in access procedure an to identify neighbor cell in handover. Cell identification technique unceasingly develops along with the evolution of cellular mobile communication system:From the frequency reuse technique of the second generation GSM system, to the planning of frequency and scrambling code in the third generation TD-SCDMA system, as well as scrambling code classifying configuration in WCDMA system, and then to use physical cell identity to achieve adjacent cell ID "conflict-free" and "confused-free" in the fourth generation system. Due to different system features, cell identification technique has various characteristics, but the core idea of resource reusing remains unchanged.As SON is the research background, the self-configuration and self-optimization of PCI and NCL are studied in-depth and comprehensive. According to the cell identification strategy of previous generations of the cellular mobile communication system, optimal allocation algorithm of PCI and NCL are proposed. Comparing with the existing techniques, feature of physical resources, resource allocation, networking performance and user experience are comprehensivly studied. The main works and innovations of the paper are:1. Rearch on properties of physical cell identityFirstly, the application of PCI in the cell search and synchronization process has been studied, including the relationship of the physical cell identity and the synchronization sequence, the generation of synchronization sequence and its properties. The paper focused on the correlation properties of Primary Synchronization Sequence (PSS) and the Secondary Synchronization Sequence (SSS), analyses the PSS and SSS properties with time and frequency offset. Research discovered that SSS are impact by the scrambling code in the time domain; therefore, there exist a set of high cross-correlation sequence. The paper analysed distribution characteristic of the high cross-correlation SSS, and a grouping allocation mechanism has been proposed.2. Rearch on self-organizing techniques of PCI in single layer ScenarioThis paper has studied on the self-organizing techniques of PCI in single layer Scenario in LTE-A system. Considering the existing algorithms only meets the basic configuration of "conflict-free"and"confusion-free", but not the pursuit of optimization objective for reuse distance. Therefore, this paper has adopted the minimum spanning tree algorithm to map the PCI allocation problem to graph theory. The greedy algorithm is used to search the maximum interference in the network, and avoided by the orthogonal PCI resources; the greedy algorithm is also used in searching the optimal solution while reusing PCI in order to minimum the multiplexing interference. This paper also presents a PCI reuse frequency based algorithm, the allocation order is determined by calculating the network adjacency matrix, and the PCI with minimum reuse frequency will be allocated to the eNB in order, to even the PCI deployment. This paper analyses the algorithm computational complexity and performance. The simulation results indicate that:the proposed algorithm with the self-configuration algorithm has lower complexity and higher user of Carrier to Interference Rate (CIR) comparing the existing self-configuration algorithm.3. Rearch on self-organizing techniques of PCI in multi-layer heterogeneous network ScenarioAccording to the characteristics of the node deployment in multi-layer heterogeneous network, the paper proposes a PCI Color Switching Algorithm (CSA) and a PCI Cassifying Configuration Algorithm (CCA) in order to adapt to the densely deployed network environment with multiple types of nodes. CSA swapped or changed the cell ID according to the local greedy searching algorithm, and then made the decision according to the cost function; the network interference is reduced gradually. CCA used the mechanism of IP address allocation to group and label the PCI resource. Network nodes can address its ID rapidly according to the adjacency cell ID. The simulation results show that the two proposed algorithms can not only guarantee the "collision-free" and "confusion-free" between the macrocells and heterogeneous nodes, but also reduce the network reuse interference effectively. 4. Rearch on self-organizing techniques of NCLThe self-organization of NCL is studied in this paper. The paper proposes a self-configuration algorithm of NCL based on signal propagation prediction, and a self-optimization algorithm for adding and removing neighbor cell caused by dynamic cell coverage is also provided. The cell measurement report and handover information are submitted by the UEs on the edge of cell. The self-optimazation of NCL is determined by the predefined threshold and the offset Time To Live (TTL), thus to achieve the dynamic changes of NCL. Through the dynamic system simulation platform, the Automatic Neighbor Relation (ANR) algorithms are simulated. The results show that after the self-organization of NCL, the Radio Link Failure (RLF) rate is improved significantly, while the success handover rate performance gain is8%.更多还原