摘要
Afuzzyneuralnetworkcontrollerforunderwatervehicleshasmanyparametersdifficulttotunemanually.Toreducethenumerousworkandsubjectiveuncertaintiesinmanualadjustments,ahybridparticleswarmoptimization(HPSO)algorithmbasedonimmunetheoryandnonlineardecreasinginertiaweight(NDIW)strategyisproposed.OwingtotherestraintfactorandNDIWstrategy,anHPSOalgorithmcaneffectivelypreventprematureconvergenceandkeepbalancebetweenglobalandlocalsearchingabilities.Meanwhile,thealgorithmmaintainstheabilityofhandlingmultimodalandmultidimensionalproblems.TheHPSOalgorithmhasthefastestconvergencevelocityandfindsthebestsolutionscomparedtoGA,IGA,andbasicPSOalgorithminsimulationexperiments.ExperimentalresultsontheAUVsimulationplatformshowthatHPSO-basedcontrollersperformwellandhavestrongabilitiesagainstcurrentdisturbance.ItcanthusbeconcludedthattheproposedalgorithmisfeasibleforapplicationtoAUVs.
出版日期
2008年03月13日（中国Betway体育网页登陆平台首次上网日期，不代表论文的发表时间）
