基于众包模式的手机基站故障诊断与性能评测

发布时间：2018-04-19 09:01

本文选题：众包 + 基站　；参考：《兰州理工大学》2017年硕士论文

【摘要】：信息技术的持续进步带动了通信服务和智能手机的迅速发展,以无线方式接入网络、功能齐全且便于携带的智能手机已成为人们处理通信、上网、娱乐和办公等事务的首选终端,进一步促进了移动通信的发展。为了满足用户需求,实现手机信号的无盲区覆盖,移动通信运营商在各地部署了大量的手机基站。基站数目增加,随之而来的是基站故障的增加以及基站带宽资源分配不合理。运营商主要通过远程监控、用户反馈以及人工巡检定位的方式来定位故障基站。但远程监控只能查看基站硬件的损害,无法查看出软件的故障;用户反馈的方式描述不详细;人工巡检会增加工作人员工作量,巡检周期较长。这些方式均无法满足现实应用的需求。运营商对基站带宽分配的方式通常是首先按照经验分配,然后根据用户的投诉进行调整。当用户投诉信息少,覆盖面不够宽广时,会缺乏对基站性能的客观评价,此时运营商迫切希望能够掌握基站性能全面且真实的反馈,尽量减少基站带宽资源过剩或者带宽资源不足的情况,真正实现资源按需分配。因此,如何高效的利用用户信息,及时而又准确的定位故障基站,客观的评价基站性能成为通信领域亟需解决的问题。本文通过分析手机切换基站的原理,提出一种廉价且高效的基于众包模式的手机基站故障诊断方法。该方法以用户手机为数据采集工具,通过数据的汇总、对比与分析,实现故障基站的定位。仿真实验表明该方法能够及时而又准确的定位故障基站。同时本文根据众包的特点,选择利用人工神经网络评价法对基站性能进行评价,结合用户评价以及智能手机采集的数据的汇聚分析,对基站性能进行排名。实验表明该方法可以实现资源的有效利用,准确而又客观的评价基站性能。本文的工作包括以下三个方面:(1)研究如何利用Android平台实现基站故障诊断与性能评测相关数据的采集,其中包括:Android平台抓包、网速测量、Ping功能实现、网络模式判别、基站信息获取。(2)在运营商制定的手机切换基站准则的基础上,提出一种定位故障基站的新方法。该方法首先根据基站分布,利用泰森多边形将城市划为若干个区域,在基站正常工作的情况下,基站信号所覆盖区域内的智能手机会自动连接该基站。一旦基站发生故障,智能手机会自动连接覆盖该区域的其他基站,通过记录这些用户的数据即可获取故障基站的位置。(3)分析众包的特点,将人工神经网络算法应用到基站性能评估中。首先利用智能手机采集一些能够反映网络性能的数据,同时记录对应的用户评分,然后采用人工神经网络模型训练该数据,根据再次采集的网络性能数据,预测出用户评分。最后通过采集得到的用户评分以及预测得到的用户评分对基站性能进行综合评价。最后,通过仿真实验验证了本文所提定位故障基站方法的可行性,以及实验表明利用人工神经网络对基站性能评价具有良好的实用性。
[Abstract]:Continuous improvement of information technology led to the rapid development of communication services and intelligent mobile phone, with wireless access network, intelligent mobile phone functions and portability has become people to deal with communication, Internet, preferred terminal entertainment and office affairs, to further promote the development of mobile communication. In order to meet the needs of users, to achieve non blind mobile phone signal coverage, mobile communication operators to deploy a large number of mobile phone base stations around the world. The number of base stations increased, followed by a base station fault and the increase of base station bandwidth resource allocation is not reasonable. Operators mainly through remote monitoring, user feedback and manual inspection positioning to locate the fault. But the base station remote monitoring base station hardware can only view the damage to look for software fault; user feedback is described in detail; the manual inspection work will increase the workload, patrol Check the cycle longer. These methods are unable to meet the needs of practical application. Operators of the base station bandwidth allocation method is usually the first according to the empirical distribution, then adjusted according to user complaints. When the user complaints less information, the coverage is not wide enough, the lack of objective evaluation of the base station, the operators are eager to master station performance feedback comprehensive and true, as far as possible to reduce the base excess bandwidth resources or insufficient bandwidth resources, realize resource allocation on demand. Therefore, how to efficiently use the user information, timely and accurate fault location of base station, the base station performance evaluation become the field of communication problems objectively. By analyzing the principle of mobile phone base station switching, provides a cheap and efficient mobile phone base station fault diagnosis method based on Crowdsourcing model. The method to user machine The data collection tool, through data collection, comparison and analysis, realize the fault location of the base station. The simulation shows that this method can locate the fault station timely and accurate. At the same time, according to the characteristics of Crowdsourcing, choose to use artificial neural network evaluation method to evaluate the performance of the base station, combined with the analysis of user evaluation and acquisition of intelligent mobile phone together the data, to rank the performance of the base station. The effective use of experiment shows that this method can realize the resources, evaluation of the performance of the base station accurately and objectively. The work of this paper includes the following three aspects: (1) research on how to realize the base station fault diagnosis and performance evaluation of data acquisition, using Android platform Android platform capture, including: speed measurement, Ping function realization, network mode discrimination, base station information. (2) based on the criterion of mobile phone base station switching operators to develop on the. A new method for fault location of base station. Firstly, according to the base station distribution, using Tyson polygons will be divided into several areas of the city, at the base station under normal working conditions, the base station signal coverage area of the intelligent mobile phone will automatically connect to the base station. The base station once a fault occurs, the intelligent mobile phone will automatically connect to other base station coverage the area of the base station through acquiring fault record these user data can be located. (3) analysis of the characteristics of Crowdsourcing, the application of artificial neural network algorithm to the base station performance evaluation. First use of intelligent mobile phone collection can reflect the network performance data recorded at the same time, the corresponding user score, then using the artificial neural network model the training data, according to the network performance data collected again, predict user scores. Finally, by collecting the user score and predicted user The performance of the base station is evaluated comprehensively. Finally, the feasibility of the proposed location based fault base station method is verified by simulation experiments, and experiments show that the artificial neural network has good practicability for the base station performance evaluation.

【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.5;TP183

【相似文献】