Analysis of machine operation skills using hand discrete movement

The function for supporting a human by considering his of her skill-level is required for a future advanced machine. In this study, discrete movements of operatorpsilas hand were analyzed to propose an adequate skill evaluation method that has possibility to be implementable to a machine. Using teleoperation experiment system for two radio-controlled construction machines, the training process of the operators was recorded, and performance of the machine operation, the reaching action of operatorpsilas hand, and the state transition matrix of the operation commands were investigated. From the analyses, sufficiently-high correlation between the change of the fitting error to Fittspsila law and the task-performance index was confirmed, and an increasing of the task performance and a decreasing of the inching had strong relationship.