基于神经网络的二关节机器人协调控制

《基于神经网络的二关节机器人协调控制》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、ABSTRACTSincetheappearanceoftheneuralnetworks，theindustrialenvironmenthasjumpedabigsteptowardthefuture，mostoftheprocessesthatacquiredahumansupervisinginthepast，hasnowtheirownlearningalgorithms，oneofthemostinfluencedfieldsbythisistheRobots，Ithenearpastmostneuralalgorithmsde

2、velopedwereforrobotscontrol，theactualobjectiveistocreatearobotthatcouldhavetheagilityofhuman，andthiscouldbeaccomplishedinthenearfutureduetothefastdevelopingofmicroprocessors，andtheappearanceofsuppercomputers．Oneofthemostchallengingapplicationinthisfieldisthevisualcontrol，k

3、nowingthatit’Sdoneinrealtime，makeitevenmoredifficultthananyoftheothermethod，mostofthesealgorithmsusetwovisualsensorstoprovidethesystemwitha3Dscene，thatcontainsthenecessarydatatOsuchasthepositionoftheobjectsandtheirshape，whichisidenticaltotheworkofhehumaneyes．InthisthesiSwe

4、willdiscussandrealizeonemethodthatusestheneuralnetworktogeneratethetasksthattherobotshoulddo．InthefirstchapterWepresentageneralview0ntheapplicationofneuralnetworksforcontrollingrobotscorn，andasthecontrolisvisualWeneedtodevelopanimageprocessingalgorithmtoreceivethenecessary

5、datafromthesensors，andthisistheSUbjectofthesecondchapter．Thechapter3discusstheneuralnetworkstheirprinciplesandsomeoftheirmostimportanttrainingalgorithms．Inchapter4wediscusstheproblemofcouplingthevisualsystemwiththerobot，andimplementthecommandsystemthatwepropose，thenextpart

6、wastrulythemostinteresting，itwassomekindofchallenge，thechapter5discusstherealizationoftwo4degreesoffreedomrobotarms，thatcouldbecontrolledfromPCwhichmakeiteasiertoimplementthecontrollerwewant，andthelastpartisthediscusstheresultsofourdesign，andtakeafinalconclusion，onthesyste

7、m，whiletheappendicesAcontainseveralissuesinprogrammingwiththevisualc十+。Bpresentthehardwareissues，suchasthesensorscalibration．．．etc，Cenclosesthelistofthetablesandthefigures，whileDcontainsthebibliography。．I．KeyWords：RobotManipulator；CooperativeControl；MachineVision；AdaptiveC

8、ontrol；IntelligentContr01．Ⅱ．湖南大学学位论文原创性声明本人郑重声明：所里交的论文是本人在导师的指导下独立进行研究所取得的研究成果。除了文中特别加以标注