基于神经网络的TD-LTE网络故障诊断技术研究

发布时间：2018-01-23 19:00

本文关键词： TD-LTE网络故障诊断 BP神经网络诊断系统　出处：《宁波大学》2014年硕士论文　论文类型：学位论文

【摘要】：TD-LTE（Time Division Long TermEvaluation）网络是TD模式的3G长期演进型网络，是由中国主导的新一代移动通信网络。与3G网络相比，它有着更高的上下行峰值速率，更高效的频谱资源利用率，更低的系统延时，在网络性能上有了质的提升，给用户带来了更好的体验感。但同时，一旦网络出现故障将会给用户带来更为明显的影响，这就要求网络运营商在网络出现故障时，能够快速有效的解决网络故障，迅速优化网络性能。而传统的故障诊断方法，其工作量大，诊断周期较长，很难实现故障的快速诊断。因此，研究快速、智能化的TD-LTE网络故障诊断技术就显得很有必要了。作为人工智能技术之一的神经网络，它有着很强的非线性处理能力，是目前实现复杂系统故障诊断智能化的一种常用技术。本文将神经网络技术引入到TD-LTE网络的故障诊断当中，研究了基于神经网络的TD-LTE网络故障诊断技术。研究工作和主要内容分为以下几个方面： 1.对TD-LTE网络及常见的一些智能故障诊断方法进行了介绍。 2.对BP神经网络基本理论和方法进行了介绍和分析，在此基础上结合网络KPI数据的特点，提出了基于KPI统计分布偏离度的BP神经网络故障诊断方法。该方法首先采集诊断所需KPI的正常历史数据，统计得到KPI的经验分布，通过KPI属性学习算法，生成KPI的属性集。然后对网络当前的KPI数据进行监测，通过异常检测方法，检测KPI的异常情况。在KPI出现异常的状况下，调用训练好的BP神经网络进行故障诊断，给出诊断结果。最后通过仿真实验证明了该方法的可行性和有效性。 3.分析和设计了基于上述诊断方法的TD-LTE网络故障诊断系统，并通过C#技术和SQLServer2008数据库实现了该系统。测试结果表明该系统可以实现网络故障的快速化、智能化诊断，验证了方案的可行性和可实现性，同时也进一步验证了本文提出的故障诊断方法的可行性和有效性。
[Abstract]:The TD-LTE(Time Division Long term value) network is a 3G long-evolving network based on TD mode. It is a new generation of mobile communication network dominated by China. Compared with 3G network, it has higher peak and downlink rate, more efficient spectrum resource efficiency and lower system delay. In the network performance has the qualitative enhancement, has brought the better experience feeling to the user, but at the same time, once the network has the breakdown will bring to the user more obvious influence. This requires network operators to solve the network failures quickly and effectively, and optimize the network performance quickly. However, the traditional fault diagnosis method has a large workload and a long diagnosis period. It is difficult to realize the fast fault diagnosis, so it is necessary to study the fast and intelligent TD-LTE network fault diagnosis technology. As one of artificial intelligence technology, neural network has strong nonlinear processing ability. It is a common technology to realize intelligent fault diagnosis of complex system. In this paper, neural network technology is introduced into fault diagnosis of TD-LTE network. The TD-LTE network fault diagnosis technology based on neural network is studied. The research work and main contents are divided into the following aspects: 1. The TD-LTE network and some common intelligent fault diagnosis methods are introduced. 2. The basic theory and method of BP neural network are introduced and analyzed, and the characteristics of network KPI data are combined. A BP neural network fault diagnosis method based on the deviation degree of KPI statistical distribution is proposed. Firstly, the normal historical data of KPI for diagnosis are collected and the empirical distribution of KPI is obtained statistically. The KPI attribute learning algorithm is used to generate the attribute set of KPI. Then the current KPI data of the network are monitored and the method of anomaly detection is used. The abnormal condition of KPI is detected. In the case of abnormal KPI, the trained BP neural network is called for fault diagnosis. Finally, the feasibility and effectiveness of the method are proved by simulation experiments. 3. The TD-LTE network fault diagnosis system based on the above diagnosis method is analyzed and designed. The system is realized by C # technology and SQLServer2008 database. The test results show that the system can realize the rapid and intelligent diagnosis of network faults. The feasibility and realizability of the proposed scheme are verified, and the feasibility and effectiveness of the proposed fault diagnosis method are further verified.
【学位授予单位】：宁波大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.5

【参考文献】