射频指标自动测试及故障诊断系统设计与实现

发布时间：2018-05-29 20:13

本文选题：射频指标 + 自动测试　；参考：《南京理工大学》2017年硕士论文

【摘要】：随着测试技术以及人工智能的发展,自动测试技术在逐步摆脱专业定制向模块化、通用化发展的同时,诊断技术也进入了以知识处理为核心,信号处理、建模处理与知识处理相结合的智能诊断技术阶段,自动测试与故障诊断的结合也一直是研究热点之一。无线通讯设备研发阶段射频指标测试当前主要采用测试人员全程值守的半自动化测试方案,然后由测试专家人工完成后期的数据分析以及测试问题定位。上述工作的效率并不能完全满足当前研发任务的需求。本课题针对测试及故障定位效率不足等问题,设计并实现了一种射频指标自动测试及智能故障诊断系统。首先对ZigBee、Sub-GHz射频指标测试需求进行了详细分析,设计了一种基于虚拟仪器技术的自动测试方案,采用华为公司Impeller环境作为测试程序运行管理平台。针对自动测试用例设计,采用测试套形式对测试仪器控制、被测设备指令输入、文档读写等通用代码进行管理,使系统具有良好的兼容性和可移植性。然后针对峰值功率、频谱模板、接收灵敏度等38个测试项的不同测试方法,分别详细说明了测试程序的实现过程,给出了相应的程序流程图。接着讨论了当前理论上较为成熟的专家系统和BP神经网络在射频指标测试问题智能定位方面的应用,设计并实现了一种以专家系统为主,BP神经网络为补充的智能故障诊断系统架构。采用面向对象技术和产生式规则结合的知识表示方法进行系统知识库设计,并使用MySQL数据库建立和维护知识库;采用基于知识的专家系统作为诊断推理核心,当故障信息与显式知识库无法匹配时,系统将调用基于数值计算的神经网络进行推理;基于预置文本法设计了解释机制,可以在诊断结果输出的同时,对诊断结果及推理过程做必要的解释;设计并实现了WEB化交互界面,在保证系统诊断性能及稳定运行的前提下,实现了友好的人机交互方式。本系统已经在华为企业网络硬件实验室(南京)投入运行,显著提升了射频指标测试效率,同时也较好实现了测试问题的智能定位,满足了设计性能需求。
[Abstract]:With the development of testing technology and artificial intelligence, automatic test technology has gradually moved away from professional customization to modularization, and at the same time, diagnosis technology has also entered the core of knowledge processing and signal processing. The combination of automatic testing and fault diagnosis has been one of the hotspots in the intelligent diagnosis technology stage which combines modeling and knowledge processing. At present, the radio frequency index test of wireless communication equipment development stage mainly adopts the semi-automatic test scheme which the testers are on duty in the whole process, and then the later data analysis and test problem orientation are manually completed by the test experts. The efficiency of the above work does not fully meet the needs of the current R & D task. In order to solve the problem of low efficiency of test and fault location, an automatic testing and intelligent fault diagnosis system for RF index is designed and implemented in this paper. Firstly, the requirements of ZigBeeg Sub-GHz radio frequency index testing are analyzed in detail, and an automatic test scheme based on virtual instrument technology is designed. Huawei's Impeller environment is used as the running management platform of the test program. For the design of automatic test cases, the test suite is used to manage the general code such as the control of test instrument, the input of instruction, the reading and writing of documents, etc., which makes the system have good compatibility and portability. According to the different testing methods of 38 test items, such as peak power, spectrum template and receiving sensitivity, the implementation process of the test program is explained in detail, and the corresponding program flow chart is given. Then it discusses the application of expert system and BP neural network, which are mature in theory, in the intelligent localization of radiofrequency index testing problem. An intelligent fault diagnosis system architecture supplemented by expert system and BP neural network is designed and implemented. The knowledge base is designed by using object-oriented technology and knowledge representation method of production rule, and the knowledge base is established and maintained by MySQL database, and the expert system based on knowledge is used as the core of diagnosis reasoning. When the fault information and the explicit knowledge base can not match, the system will call the neural network based on the numerical calculation for reasoning, and design the explanation mechanism based on the preset text method, which can output the diagnosis results at the same time. The WEB interactive interface is designed and implemented, and the friendly man-machine interaction mode is realized on the premise of ensuring the diagnostic performance and stable operation of the system. The system has been put into operation in the Huawei Enterprise Network hardware Laboratory (Nanjing), which has greatly improved the efficiency of RF index testing, at the same time, it has realized the intelligent positioning of the test problem and satisfied the requirement of design performance.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP274;TP277

【参考文献】