【作者】 李兵；

【导师】 何怡刚；

【作者基本信息】 湖南大学 ， 电气工程， 2011， 博士

【摘要】 随着近几年“物联网”技术的飞速发展,作为物联网关键技术之一的射频识别技术(Radio Frequency Identification, RFID)技术也迎来了一个新的飞速发展时期,开始大量应用于工业生产自动化、交通、身份识别和物流等领域,并且在不断扩大其应用范围。同时,RFID技术标准的多样性及应用的特殊性给相应的RFID测试和RFID性能评估也带来新的要求和挑战。基于以上现状,应用电磁场、天线、测试和统计分析等相关理论,本文研究了无源超高频RFID系统中的雷达散射截面差值、误码率和反向散射调制性能以及RFID系统性能评估模型,主要内容及工作包括：首先阐述了RFID系统工作原理、组成及RFID相关标准。从RFID测试原理、测试技术和测试内容出发,根据应用测试现状研究了RFID系统性能测试方法及测试设备；介绍了虚拟仪器技术、软件无线电技术和射频测试技术,基于此设计了一套基于虚拟仪器技术和软件无线电技术的RFID系统测试平台,重点介绍了测试系统的架构形式、硬件选型及软件设计,并给出了一些具体参数的测试方法和测试结果。结果表明,该测试平台具备可靠性、兼容性、稳定性等性能测试能力,可以完成RFID空中接口一致性、电气性能、第三方监测、系统性能及应用场景性能等测试,具备性能指标分析和可视化性能评估功能,并且支持多协议标准的测试,具备客户自定义标准测试的扩展能力。可以满足RFID术研究测试、硬件产品认证测试、应用测试等公共服务需求。其次基于RFID技术通信原理和天线反向散射理论,导出了无源标签的雷达散射截面的三维表现形式,并给出了反向散射信号在“0”“1”状态下的雷达散射截面差值与反向散射链路通信误码率的关系,提出了一种适合于反向散射RFID通信链路的误码率分析及测试方法。使用研发的RFID测试系统,运用矢量相减法消除收发天线间直接路径泄漏信号产生的误差,测量了不同参数设置情况下标签的雷达截面差值和通信误码率。得出在满足阅读器和标签灵敏度要求条件下,发射功率越大,标签反向散射链路通信BER越大。第三,分析了无源超高频射频识别系统阅读器接收机取得最大有效吸收功率值的条件,讨论阻抗失配对标签反向散射链路调制系数的影响,导出阅读器接收机标准化有效吸收功率、解调输出信号信噪比(SNR)下边界和接收端误码率(BER)三者的反向散射调制系数表达式。在开阔的室内环境下,完成了不同参数条件下的反向散射调制系数测试。采用研发的RFID测试系统对采用ASK调制方式的标签调制系数进行了测试,得出：调制系数位于5%～10%区间时,标签可以被正确识别；调制系数大于10%时,芯片吸收功率过低而无法正常工作,标签反向调制信号不能被正确解调和解码；调制系数小于5%时,阅读器接收机有效吸收功率过低,接收端BER过高,标签无法被识别。最后,基于实验设计方法理论,设计了一个RFID系统性能测试实验,利用统计分析方法给出了影响RFID系统性能的主要因素,并建立了RFID系统性能评估回归模型。基于OPC技术的Lab VIEW与PLC实时通讯控制技术,构建了一个仿真RFID系统应用场景,提高了测试效率,减少了测试成本。结合LVQ算法和GA算法,提出了一种改进的GA-LVQ算法,并将其引入RFID系统标签识别率预测,一方面利用LVQ神经网络强大的学习、联想、计算简单和分类识别的功能,并引入学习速度修正规则,保证算法的学习稳定性,另一方面利用GA算法较强的全局搜索功能、求解能力和鲁棒性,并引入交叉率和变异率算子动态调整方法,避免“优秀基因丢失”和“种群单调”问题,提高算法的学习效率。实验结果表明,该算法具有较好的预测精度、学习效率和鲁棒性。更多还原

【Abstract】 In recent years, with the rapid development of internet of things (IOT), radio frequency identification (RFID) as one of the key technologies also achieved a new rapid development. RFID has a comprehensive application in many fields, such as industrial automation, transportation, identification and logistics etc, and expands its application scope continually. Meanwhile, the diversity of RFID standards and the particularity of the application also bring new demands and challenges to the test and the performance evaluation of RFID systems. Based on these, the delta radar cross section (delta RCS), bit error rate (BER), backscattered modulation and performance evaluation model of passive ultra high frequency (UHF) RFID systems are researched in this paper, which applies with electromagnetic fields, microwave and antennas testing and statistical analysis method. The main content and work are as follows:Firstly, this paper describes the basic theory, compositions and relevant standards of RFID systems. From the RFID testing principles, technologies and contents, the test methods and equipment of RFID system are researched. Then a test platform is built based on the virtual instrument technology, software defined radio (SDR) and RF test technologies. This paper introduces the key of the architectural form, the selection of hardware and the design of software, and gives some measuring methods and results for specific parameters. The results show that the platform has ability for test of reliability, compatibility and stability, which can complete the test of protocol and conformance, electrical properties, sniffer, and performance of system and application scenarios. It possesses the function of performance index analysis and visualization of performance evaluations, which supports multi-protocol and user-defined standards. It meets demand of public services such as RFID technology research and test, hardware product certification test, application test.Secondly, based on theory of RFID technology and backscatter, a relationship is derived between BER of backscatter link and the passive tag’s delta RCS of state "0" and state "1". Therefore, this paper presents a new method which is adapted to analysis and test for BER of RFID communication link. A measurement system based on virtual instruments is built, and introduces vector subtraction methods to eliminate the errors brought by direct path leakage signal between transmit and receive antennas. Furthermore delta RCS and the BER of backscattered signal under different parameters are measured and analyzed. Experimental results show that the greater the transmitted power, the bigger the BER of backscatter link when the received power meet required sensitivity of readers and tags.Thirdly, optimal conditions for maximizing the effective received power of reader receiver for passive UHF RFID system are analyzed in this paper. The modulation index of tag backscatter link which affectied by mismatched impendence is discussed. Representations based on backscattered modulation index are derived, which are used for calculating normalized effective received power of reader receiver, lower boundary of signal noise ratio (SNR) for demodulated output signal and BER. Modulation index of backscatter link under different parameters is measured in open indoor environment. The measurement results show that the tag can be detected successfully when the modulation index of backscatter link is in 5% to 10%. When the modulation index of backscatter link is greater than 10 percent, the power absorbed by chip of tags is too low to work normally; the modulated signal of backscatter link cannot be demodulated and decoded correctly. When the modulation index is less than 5 percent, the effective power received by the receiver of readers is too low and the BER is too high, so the tags cannot be identified successfully.Finally, an experiment is designed to test the performance of RFID systems based on the theory of design of experiment (DOE). The RFID systems performance evaluation regression model is established with the key influence factors affecting the performance of RFID system which provided by statistical analysis method. The simulated applications scenarios are built to enhance the test efficiency and reduce the cost, which is based on the OPC technology of LabVIEW and real time control technology of PLC. An improved algorithm based on learning vector quantization (LVQ) and genetic algorithm (GA) is introduced in the prediction of the identification rate of RFID systems. The GA-LVQ algorithm takes advantage of the powerful learning, association, simple calculation and classification of LVQ and strong global search function, solving ability and robustness of GA. Meanwhile, a modified learning speed rule is introduced to guarantee the stability of learning algorithm; on the other hand, a dynamic adjustment method of crossover and mutation is introduced to avoid loss of excellent gene and population drab problems, and improve the learning efficiency. Experimental results show that the GA-LVQ algorithm has better precision, learning efficiency and robustness.更多还原