Abstract: SOM (Self-organizing Maps) model was introduced†† to†† cluster†† and†† analyse on†† the†† human†† grasping activities of GloveMAPbasedondatareductionoftheinitial grasping data. By acquiring the data reduction of the initial hand graspingdataoftheseveralobjects,itwillbegoingtobe functionedastheinputstotheSOMmodel. Aftertheiterative learningofnet-trained,alldataofthetrainednetworkwillbe simulated†† and†† finally†† self-organized.†† The†† output†† results†† of modelsí are farthest approached to the reality in 3-dimensional grasping features.Theexperimentalresultofthesimulation signalwillgeneratethesimulateresultofthegraspingfeatures fromtheselectedobject.The whole experiment of grasping features is derived into three features / groups and the results are satisfactory.

Keywords: Cluster†† Analysis, Data†† reduction, Fingers grasping, Grasping features, SOM neural networks