摘要 : On manipulation of usual machines, adequate selection of operational commands and optimal scheduling of tasks are significant for skilled operation. In the present research, an analysis procedure to make the task switching charact... 展开 On manipulation of usual machines, adequate selection of operational commands and optimal scheduling of tasks are significant for skilled operation. In the present research, an analysis procedure to make the task switching characteristics visible is presented by utilising self-organising map (SOM). For experimental verification, a remote operation system was built that simulates the operation of actual construction equipments, and the process of learning its operation was analysed. It was confirmed that the proposed method was effective through investigation of matching ratios between commands predicted by the trained SOM and actual ones. In addition, a transition analysis method of the learning processes is presented using geometrical distances and areas made by best matching nodes on an SOM planar map. This method was valid to an early phase SOM where clusters are not formed due to inexperienced feature of the operator. Several characteristics concerning the learning process were found, and an ability to show differences of the user skill level was confirmed. 收起