Visualization of task switching strategy of machine operation

For a future advanced machine, functions for supporting human by considering not only the situation around the machine but also the operator´s skill-level are required. Therefore, an establishment of a method to evaluate the skill is expected. On the general machine manipulation, the task scheduling and the adequate selecting of operation commands are important to execute sequential task successfully. In the present research, an analysis procedure to make the task-switching strategy visible is presented by utilizing the self-organizing map (SOM). For practical verification, the remote operation experiment system was built with radio-controlled construction machines, and the proposed analysis procedure was applied to the experimental task. As the result, the matching ratio of the command predicted by the trained SOM was sufficiently-large, and the effectiveness of the proposed method was confirmed.