资源描述:
《A three-dimensional proportional guidance law based on RBF neural network》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、Proceedingsofthe7thWorldCongressonIntelligentControlandAutomationJune25-27,2008,Chongqing,ChinaAThree-dimensionalProportionalGuidanceLawBasedonRBFNeuralNetworkWenjinGuHongchaoZhaoRuchuanZhangNo.3DepartmentFaculty703No.3DepartmentNavalAeronauticalandAstronauticalNavalAeronauticala
2、ndAstronauticalNavalAeronauticalandAstronauticalUniversityUniversityUniversityYantai,ShandongProvince,ChinaYantai,ShandongProvince,ChinaYantai,ShandongProvince,Chinafxqyjs@sina.comnavyzhc@163.comfxqyjs@sina.comAbstract-Anewthree-dimensional(3D)pureproportionalsolutionstothedynami
3、coptimizationproblemexist.Thus,inguidancelawbasedonradialbasisfunction(RBF)neuralmostreal-lifeapplications,thefirst-orderOGL(Cottrel,1971)network(PNGLRBF)designmethodwasproposed.Theisused,resultinginareducedrobustnesstomodelandproposedmethodwasappliedtodecreasingthemissdistance,[
4、1]parameteruncertainties(Weiss&Hexner,1998).whichismostlyarisenfromthefixednavigationratesofThepureproportionalnavigationguidance(PPNG)lawstraditionalproportionalnavigationguidancelaws(TPNGLs).hasbeenwidelyrecognizedasoneofthemostefficientandFirstly,usingtheengagementgeometryform
5、ulabetweenmissileandaircraft,atypical3Daugmentproportionalguidancelawpracticalguidancelaws.Forthecaseofnon-maneuveringwaspresented.Secondly,asimpleperformanceindexaboutmissaircrafts,Becket[2]derivedaclosed-formsolutionofthedistancewascreated,andthenRBFneuralnetworksweredynamicequ
6、ationsfor2-dimensionalpursuitsituationsandemployedtotunethenavigationratesonlinetoproduceoptimalanalyzedthecaptureabilityofthePPNGlawsutilizingtheover-loadcommands.Finally,twoengagementscenariosweresolution.Forthecaseofmaneuveringaircrafts,Guelman[3]examined,andacomparisonbetween
7、TPNGLsandtheproposedusedthequalitativeanalysismethodtoanalyzethecapturePNGLRBFwasmade,thesimulationresultsshowthattheabilityofthePPNGlawsagainstaaircraftmaneuveringwithproposedtheproposed3DPNGLRBFcanachieveidealmissconstantnormalacceleration.WhilethesimpleandeasytodistancethanTPN
8、GLs.implementguidancelaws,thePN,doesnotp