資源描述:
《基于時(shí)延補(bǔ)時(shí)償算法的網(wǎng)絡(luò)控制系統(tǒng)研究》由會(huì)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、摘要摘要網(wǎng)絡(luò)控制系統(tǒng)是復(fù)雜控制系統(tǒng)和遠(yuǎn)程控制系統(tǒng)的客觀需求,與傳統(tǒng)的點(diǎn)對(duì)點(diǎn)的控制系統(tǒng)相比,由于其易于擴(kuò)展和維護(hù)、可靠性高和易于實(shí)現(xiàn)資源共享等優(yōu)點(diǎn),具有廣泛的應(yīng)用前景,從而成為當(dāng)今控制學(xué)界研究的熱點(diǎn)。網(wǎng)絡(luò)是一把雙刃劍,在帶來了諸多便利的同時(shí)也帶來了諸多問題,比如:網(wǎng)絡(luò)時(shí)延、數(shù)據(jù)包丟失和網(wǎng)絡(luò)調(diào)度等問題,這些問題給控制系統(tǒng)的設(shè)計(jì)帶來了很大的困難。本論文在分析網(wǎng)絡(luò)控制系統(tǒng)的特點(diǎn)和國內(nèi)外研究現(xiàn)狀的基礎(chǔ)上,把網(wǎng)絡(luò)時(shí)延問題作為主要的研究方向。針對(duì)網(wǎng)絡(luò)時(shí)延問題,本論文主要做了以下兩個(gè)方面的工作:1.時(shí)延估計(jì)。首先在大量實(shí)測時(shí)延的基礎(chǔ)上,建立了網(wǎng)絡(luò)時(shí)延的自適應(yīng)AR模型,在兼顧預(yù)測精度和實(shí)時(shí)性的基礎(chǔ)上
2、,確定了模型的階數(shù),并利用最小均方算法根據(jù)新的時(shí)延值在線調(diào)整模型的加權(quán)系數(shù),通過建立的模型對(duì)時(shí)延進(jìn)行預(yù)測。并通過仿真驗(yàn)證了文中的算法對(duì)時(shí)延有很好預(yù)測效果。2.時(shí)延補(bǔ)償。在前面時(shí)延估計(jì)的基礎(chǔ)上,針對(duì)網(wǎng)絡(luò)時(shí)延的補(bǔ)償問題,本文提出了以下兩種時(shí)延補(bǔ)償算法:a.一種是改進(jìn)的快速廣義預(yù)測控制算法,避免了在線求解Diophantine方程,減少了計(jì)算時(shí)間,較好地保證了網(wǎng)絡(luò)控制系統(tǒng)的實(shí)時(shí)性;b.另一種是根據(jù)時(shí)延落在不同區(qū)間的多模型切換控制算法,當(dāng)時(shí)延小于一個(gè)采樣周期時(shí),采用算法實(shí)時(shí)性較好的LQG最優(yōu)控制算法,當(dāng)時(shí)延大于一個(gè)周期時(shí),采用改進(jìn)的快速廣義預(yù)測控制的算法。并通過仿真證明了該方法對(duì)網(wǎng)絡(luò)時(shí)延具
3、有很好的補(bǔ)償效果,且具有很好的實(shí)時(shí)性。關(guān)鍵詞:網(wǎng)絡(luò)控制系統(tǒng);網(wǎng)絡(luò)時(shí)延;時(shí)延估計(jì);切換控制大連交通大學(xué)r學(xué)碩十學(xué)位論文AbstractNetworkcontrolsystemistheobjectiverequirementsforthecomplexcontrolsystemandtheremotecontrolsystem,Comparedwiththetraditionalpointtopointcontrolsystem,ithasawiderangeofapplicationsprospectsbecauseofitseasytoextendandmaintainandea
4、setoshareofresourcesandSOon,andthusbecomethehotspotsoftheresearchofcontroltheory.Networkisadouble—edgedsword,whichmightleadsalotofconveniencealsobringsmanyproblems,suchas:networktimedelays,packetlossandnetworkschedulingandSOon.Theseproblemsmakegreatdifficultiestothedesignofnetworkcontrolsystem
5、.Inthispaper,basedontheanalysisofnetworkcontrolsystem’Scharacteristicsandtheresearchstatusofhomeandabroad,theissueofnetworktimedelaysisthemajorresearchdirectionofmypaper.Fortheissueofnetworktimedelays,thisthesismadetheworkoftwoaspectsasfollowing:1.Time-delayestimation.First,alargenumberofmeasu
6、redtimedelay,basedontheestablishmentofanetworkdelayadaptiveARmodel,inbothpredictionaccuracyandreal-timebasis,todeterminetheorderofthemodel,andusetheleastmeansquarealgorithmbasedonthenewdelayValue-lineadjustmentofthemodelweightingcoefficient,byestablishingamodeltopredictdelay.Throughthesimulati
7、onofthetimedelayestimationalgorithminthispaperhaveagoodpredictionofthedelayeffect.2.Thetimedelaycompensation.Basedontheabovetime—delayestimation,fortheissueofcompensationfornetworkdelay,thispaperpresentstwokindsoftime—delaycompensationa