扩散硅压力传感器迟滞建模与补偿研究

分类号密级UDC编号硕士学位论文题目扩散硅压力传感器迟滞建模与补偿研究（英文）StudyonModelingandCompensationforHysteresisofDiffusedSiliconPressureSensor独创性（或创新性）声明本人声明所呈交的论文是我个人在导师指导下进行的研究工作及取得的研究成果。尽我所知，除了文中特别加以标注和致谢中所罗列的内容以外，论文中不包含其他人已经发表或撰写过的研究成果；也不包含为获得桂林电子科技大学或其它教育机构的学位或证书而使用过的材料。与我一同工作的同志对本研究所做的任何贡献均已在论文中做了明确的说明并表示了谢意。申请学位论文与资料若有不实之处，本人承担一切相关责任。本人签名：日期：摘要I摘要扩散硅压力传感器是目前用于测量压力的一种重要传感器，被广泛应用于工业、农业、气象部门、航天航空等领域。随着半导体技术的发展，扩散硅压力传感器的各项技术指标不断提高，但扩散硅压力传感器本身固有的非常规、多值对应的迟滞非线性特性，限制了其测量精度。目前国内外针对传感器迟滞特性的研究还比较少，已有文献证明不能直接使用神经网络对其进行建模，大部分学者的工作是把迟滞非线性当作一般非线性处理，其结果不能满足补偿精度的要求。因此本文引入迟滞算子的概念，建立T-S模糊动态迟滞模型，实现对迟滞的智能补偿，并在此基础上设计了基于ARM的嵌入式智能扩散硅压力传感器补偿系统，本文主要获得如下成果：1.借鉴提升输入维数的思想，引入一个迟滞算子，将迟滞非线性的多值映射转化为单值映射，建立一个基于算子的T-S模糊动态迟滞模型，该模型具有结构简单、易于辨识算法等优点。2.引入与迟滞算子作用类似的迟滞逆算子，将多值映射转化为单值映射，建立一个基于逆算子的T-S模糊动态迟滞逆模型，将该逆模型置于被补偿对象后面进行迟滞非线性补偿，实现了对扩散硅压力传感器的智能补偿。3.在32位ARM微处理器的EM9000嵌入式模块平台上，设计了WindowsCE嵌入式扩散硅压力传感器补偿系统，完成了嵌入式扩散硅压力传感器系统的设计与开发，并结合迟滞智能补偿原理，实现了对扩散硅压力传感器高精度测量的实时性要求，与未补偿的扩散硅压力传感器迟滞相比精度提高了约四倍，证明了建立的T-S模型是非常有效性的。关键词：迟滞算子，T-S模糊模型，嵌入式系统，WindowsCE，ARMAbstractIIAbstractThediffusedsiliconpressuresensorisanimportantdevicetomeasurethepressure，whichhasbeenwidelyusedinthefieldofindustry,agriculture,meteorology,aviationandsoon.Withthedevelopmentofsemiconductortechnology,theperformanceofdiffusedsiliconpressuresensorisimproving.However,Thehysteresischaracteristicofdiffusedsiliconsensorisakindofnon-smoothnon-linearitywithmult-ivaluedmapping,whichreducestheprecisionofmeasurement.Uptillnow,fewauthorsstudythehysteresisofsensorsandpublishedliteratureprovedthatneuralnetworkscan’tapproximatethehysteresisnon-linearitydirectly.Thecompensationforthehysteresisofthesensorusuallyemploystheconventionalmethods,butthemethodscan’ttotallymeettherequirmentofcompensationprecision.Therefore,ahystereticoperatorofsimplestructureisintroducedtoestablishthehysteresismodelandputforwardtomakeintelligentcompensationtothehysteresis.Hence,anembeddedsmartdiffusedsiliconpressuresensorsystembasedonARMisdesigned,andthemainresultsareasfollows:1.Thehystereticoperatorbasedontheinputextendedspacemethodisproposedtotransformthemulti-valuedmappingintoaone-to-onemapping.Inthismethod,anovelhystereticoperatorisconstructedtoextendtheinputspacetoestablishT-Sfuzzydynamichysteresismodel.Ithasarathersimplestructure,simplifiesidentificationprocedure.2.Hystereticinverseoperatorwhichhasthesimlarfunctionswithhysteresisoperator,i.e.transformingthemulti-valuedmappingintoaone-to-onemapping,isproposed.Then,T-Sfuzzydynamichysteresisinversemodelispresented,andnonlinearhysteresischaracteristicsarecompensatedbytheinversemodelbehind,whichrealizesintelligentcompensationforthehysteresis.3.Onthe32-bitsARMmicroprocessorEM9000embeddedmodule,theWindowsCEembeddedcompensationsystemaboutthediffusedsiliconpressuresensorisdesigned.Then,thesystemaboutthediffusedsiliconpressuresensorisdeveloped,whichcombineswithprincipleofthecompensation.Suchreachesthereal-timedemandofthediffusedsiliconpressuresensor,fourtimeshigherthanothermethodwithoutcompensationinprecision.ResultshowsthatT-Smodelisextraordinarilyeffective.Keywords：hysteresisoperator，T-Sfuzzymodel，embeddedsystem，WindowsCE，ARM目录III目录摘要....................................................................................................................................IABSTRACT........................................................................................................................II目录.................................................................................................................................III第一章绪论.........................................................................................................................1§1.1扩散硅压力传感器基本结构和原理介绍..............................................................1§1.1.1扩散硅压力传感器基本结构..........................................................................1§1.1.2扩散硅压力传感器基本原理..........................................................................2§1.1.3影响扩散硅压力传感器迟滞特性的因素......................................................2§1.2扩散硅压力传感器国内外发展状况.....................................................................3§1.3课题来源和主要研究内容.....................................................................................4§1.4论文结构安排.........................................................................................................4第二章迟滞的定义及数学模型介绍.................................................................................6§2.1迟滞的定义.............................................................................................................6§2.2迟滞数学模型.........................................................................................................7§2.2.1基于微分方程的模型......................................................................................8§2.2.1.1Duhem模型................................................................................................8§2.2.1.2Jiles-Atherton(JA)模型...............................................................................8§2.2.1.3Chua-Stromsmoe模型.......................................................