資源描述:
《低維混沌神經(jīng)網(wǎng)絡(luò)的電路設(shè)計》由會員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫。
1、重慶郵電大學(xué)碩士學(xué)位論文低維混沌神經(jīng)網(wǎng)絡(luò)的電路設(shè)計姓名:田巧麗申請學(xué)位級別:碩士專業(yè):微電子與固體電子學(xué)指導(dǎo)教師:楊曉松20060515重慶郵電大學(xué)碩士論文摘要混沌神經(jīng)網(wǎng)絡(luò)對于信息產(chǎn)生、模式識別、聯(lián)想記憶、自適應(yīng)學(xué)習(xí)等都具有重要意義?;煦缟窠?jīng)網(wǎng)絡(luò)的研究起源于并基于混沌神經(jīng)元的研究,混沌神經(jīng)元是構(gòu)造混沌神經(jīng)網(wǎng)絡(luò)的基本單位,構(gòu)造和討論其硬件實(shí)現(xiàn),不僅可以直接觀察混沌神經(jīng)網(wǎng)絡(luò)的動力學(xué)行為,還可以為制造混沌神經(jīng)計算機(jī)打下基礎(chǔ)。本文在所發(fā)現(xiàn)低維混沌神經(jīng)網(wǎng)絡(luò)的基礎(chǔ)上,將其數(shù)學(xué)模型轉(zhuǎn)換為電路方程,通過混沌電路設(shè)計中開關(guān)線性控制和模塊化設(shè)
2、計思路,將神經(jīng)網(wǎng)絡(luò)的基本構(gòu)成單位神經(jīng)元設(shè)計為子電路模塊,通過神經(jīng)元子電路模塊的耦合,構(gòu)造出低維混沌神經(jīng)網(wǎng)絡(luò)電路,從而對混沌神經(jīng)元及神經(jīng)網(wǎng)絡(luò)進(jìn)行模擬。以三維混沌Hopfield神經(jīng)網(wǎng)絡(luò)模型為例,詳細(xì)闡述了混沌神經(jīng)網(wǎng)絡(luò)的計算機(jī)模擬仿真,神經(jīng)元的子電路構(gòu)造,神經(jīng)元模塊之間的耦合。最后對整個系統(tǒng)合理的修改簡化,以適應(yīng)電子元器件的參數(shù)要求,并最終將其電路(或集成電路)實(shí)現(xiàn)。在此基礎(chǔ)上又分析了四維超混沌神經(jīng)網(wǎng)絡(luò)模型,沿用模塊化思路進(jìn)行擴(kuò)展,完成了電路設(shè)計。由此,依照此方法,可實(shí)現(xiàn)低維混沌神經(jīng)網(wǎng)絡(luò)電路的設(shè)計,也為以后多維混沌神經(jīng)網(wǎng)絡(luò)電路
3、設(shè)計提供了參考。這樣就大大地縮短了我們對混沌神經(jīng)網(wǎng)絡(luò)電路模擬的周期,簡化了混沌神經(jīng)網(wǎng)絡(luò)的電路設(shè)計,為我們研究混沌神經(jīng)網(wǎng)絡(luò)及混沌電路提供了便利。關(guān)鍵詞:混沌,神經(jīng)網(wǎng)絡(luò),混沌電路,神經(jīng)元模塊,吸引子,極限環(huán)重慶郵電大學(xué)碩士論文摘要AbstractChaoticneuralnetworkshaveimportantsignificanceforinformationproducing,patternrecognition,memoryrecalling,self-adaptivelearningandSOon.Theresear
4、chaboutitiSresourcedandbasedonchaoticneuronswhichiSbasedcellforcomposingneuralnetworks,SOitcanobservetheresponseofchaoticneuronsandneuralnetworksbymakinganddiscussingitshardwarecircuitIt’Snotonlydirectlyobservethedynamicsbehaviorofchaoticneuralnetworks,butalsoprep
5、aredformakingchaoticcomputersInthispaper,wepresentlow-dimensionalchaoticneuralnetworksmodels,andwetransformthemodelstocircuitequationsfirst.Usingtheideaofon—offlinearcontrolandmodularizationdesign,wedesignandbuildthelow—dimensionalchaoticneuralnetworkscircuRsbycou
6、plingtheneuralsub-circuitmodules,thenwecansimulatethebehaviorsofthechaoticneuronsandneuralnetworksTakinga3Dchaoticneuralnetworksmoduleforexample,weexpatiatethemakingofneuralsub·circuits,thecouplingbetweenthemodules,simulatingthechaoswhichcanbeeasilyprovedbythetopo
7、logicalhorseshoestheoryforthepiecewisecontinuousmapsbycomputer’Said:finally,thewholesystemismodifiedandsimplifiedtomeettherequirementsoftheelectronicdevicesanditsparameters,andimplementedwithacircuitoraintegratecircuitAndweanalyze41)chaoticneuralnetworks,designthe
8、circuitbythisway.Wecanexpandbythismethodtoimplementthecircuitdesignoflow-dimensionalchaoticneuralnetworks,anditgivesUSareferenceforhigh—dimensionalchaot