四旋翼無(wú)人機(jī)視覺導(dǎo)航與軌跡跟蹤

旋■無(wú)人機(jī)視覺導(dǎo)航與軌跡跟蹤AbstractWiththecontinuousdevelopmentofcomputerhardwaretechnologyandsoftwarealgorithms,theroleofquadrotorUAVsincivilandmilitaryapplicationsisincreasinglyimportant.Traditionalnavigationmethodsandconventionalcontrolalgorithmshavebeenunabletomeettheincreasinglycomplexrequirementsofquadrotor’smission.Therefore,thispaperwillstudythequadrotorUAVvisionnavigationalgorithmandtrajectorytrackingcontrolalgorithm.TheflightmechanismofquadrotorUAVisanalyzedinthispaperandthecoordinatesystemisdefined.ThecoordinatetransformationequationsbasedonEuleranglesandquaternionsarederivedandthequadrotor'skinematicsanddynamicsmodelareestablished.AimingattheproblemthatthecalculationofdesiredattitudecomputedfromthepositionloopbasedontheEuleranglemodelisexpectedtobecomplexandpronetosaturation,anewmethodisproposed,whichcanobtainthedesiredquaternionfromtheaccelerationobtainedbythepositionloopthroughgeometrictransformationandsimplifythequadrotorsystemmodel.ThesystemmodelequationestablishesthemodelfoundationforthedesignofthetrajectorytrackingcontrolsystemofthequadrotorUAVTheUAVvisionnavigationsystemisdesignedandthefunctionofeachpartofthesystemisintroducedatthesametime.AimingattheproblemthattheGPSnavigationmethodscannotbeaccuratelylocatedwhenthesatellitesignalisweak,atargetdetectionalgorithmbasedonL-Kopticalflowmethodisdesignedtorealizethedetectionofdynamictargets.Atthesametime,thispaperstudiesthetargettrackingalgorithm.TheprincipleofMeanshiftalgorithmisintroduced,andtheCamshifttargettrackingalgorithmisimplementedbasedonthis.FortheproblemthatthetargettrackingiseasilylostbasedonthetraditionalCamshiftalgorithminthecompliexenvironmentbackground,theCamshiftalgorithmfusingtheKalmanfilterisproposed,whichvoidstheproblemoftrackingfailurewhenthetargetisoccludedtemporarily,andimprovestherobustnessoftargettrackingsystem.AimingattheproblemthattheconventionalcontrolalgorithmcannotmeettherobustnessofthenonlinearmodelcontrolofthequadrotorUAV,thispapermakesadeepstudyontheactivedisturbancerejectioncontrol(ADRC)algorithm.Firstly,thecompositionandfunctionofeachpartoftheADRCareanalyzed,andthestabilityoftheESOisproved.Then,theattitudedecouplingcontrollerbasedonADRCisproposedtorealizetherobustcontroloftheattitudeloopofthequadrotorUAV;Aimingatthecharacteristicsoftheunderactuatedandmulti-channelcouplingoftheactualquadrotorUAVmodel,thedoubleclosed-loopPIDcontrollerofpositionloopandlinaeractivedisturbancerejectioncontrollerofattitudeloopbasedonthequaternionnonlinearmodelequationaredesigned,thetrajectorytrackingcontrolsystemrealizesinterferencerejectionandtrackthereferencesignalsaccurately,whichimprovestherobustnessofthesystemandthesimulationshowstheeffectivecomparedwiththePID.Finally,thispaperestablishesthehardwareplatformandsoftwareplatformofquadrotorUAVvisionnavigationandtrajectorytrackingsystem,andsometestsarecarriedoutbasedtheactualsystem.ThequadrotorUAVtrajectorytrackingtestiscompletedbaseontheopen-sourcePixhawkflightcontrolsystem;hoveringindoorsandtargettrackingofquadrotorUAVaretestedbaseonthedesignedvisionnavigationalgorithms.Keywords:QuadrotorUAV,visionnavigation,L-KOpticalflow,Camshift,Kalman,ADRC,TrajectoryTracking目錄摘.…. IAbstract II第1章緒論..(1)1.1課題研究的背景和意義(1)1.2國(guó)內(nèi)外研究現(xiàn)狀及分析(3)1.2.1視覺檢測(cè)與跟蹤技術(shù)(3)1.2.2四旋翼無(wú)人機(jī)控制方法(5)1.3主要研究?jī)?nèi)容與結(jié)構(gòu)安排(6)1.3.1本文的研究?jī)?nèi)容(6)1.3.2本文的組織結(jié)構(gòu)(7)第2章四旋翼無(wú)人機(jī)數(shù)學(xué)建模(9)2.1引言(9)2.2四旋翼無(wú)人機(jī)工作原理分析(9)2.3四旋翼無(wú)人機(jī)的動(dòng)力學(xué)模型建立及分析(10)2.3.1坐標(biāo)系定義及姿態(tài)描述(10)2.3.2四旋翼無(wú)人機(jī)動(dòng)力學(xué)模型(11)2.4視覺導(dǎo)航系統(tǒng)設(shè)計(jì)(16)2.5本章小結(jié)(18)第3章視覺導(dǎo)航算法設(shè)計(jì)(19)引言(19)3.2基于L-K光流法的運(yùn)動(dòng)檢測(cè)(19)3.3Camshift目標(biāo)跟蹤算法(22)Meanshift算法原理(22)Camshift算法原理(26)3.3.3融合Kalman濾波的Camshift算法實(shí)現(xiàn)(29)3.3.4實(shí)驗(yàn)結(jié)果(32)3.4本章小結(jié)(34)第4章四旋翼無(wú)人機(jī)軌跡跟蹤控制系統(tǒng)設(shè)計(jì)(35)引言(35)4.2自抗擾控制器原理(35)4.3基于自抗擾的四旋翼姿態(tài)解耦控制器設(shè)計(jì)(42)4.3.1姿態(tài)自抗擾控制器設(shè)計(jì)(42)4.3.2數(shù)值仿真分析(43)4.4基于非線性模型的控制器設(shè)計(jì)(47)4.4.1位置環(huán)PD控制器設(shè)計(jì)(47)4.4.2姿態(tài)環(huán)線性自抗擾控制器設(shè)計(jì)(48)4.4.3數(shù)