%0 Journal Article
%T 对称工件定位算法:收敛性及其改进
%A 陈善勇
%A 李圣怡
%A 戴一帆
%J 自动化学报
%P 428-432
%D 2006
%X ？Symmetricworkpiecelocalizationalgorithmscombinealternatingoptimizationandlinearization.Theiterativevariablesarepartitionedintotwogroups.Thensimpleoptimizationapproachescanbeemployedforeachsubsetofvariables,whereoptimizationofconfigurationvariablesissimplifiedasalinearleast-squaresproblem(LSP).Convergenceofcurrentsymmetriclocalizationalgorithmsisdiscussedfirstly.ItisshownthatsimplytakingthesolutionoftheLSPasstartofthenextiterationmayresultindivergenceorincorrectconvergence.Thereforeinourenhancedalgorithms,linesearchisperformedalongthesolutionoftheLSPinordertofindabetterpointreducingthevalueofobjectivefunction.Wechoosethispointasstartofthenextiteration.Betterconvergenceisverifiedbynumericalsimulation.Besides,imposingboundaryconstraintsontheLSPprovestobeanotherefficientway.
%K Symmetricworkpiecelocalization
%K algorithm
%K improvement
%K linesearch
%U http://www.aas.net.cn/CN/abstract/abstract15831.shtml