부산대학교 도서관

The performance of a mobile working machine is strongly dependent on the skills of the operator. This paper describes a method for coaching operators of mobile working machines, aiming at increasing the performance of the machine work. The first step is to recognize the plan of the operator, the intent behind the operator's actions. The plan is recognized from the control signals given by the operator during normal work, using hidden Markov models. The operator's performance level is evaluated using the obtained state information. The performance of the operator in a man-machine system is regarded to consist of cognitive and motor-sensory skills. The randomness associated with human operators is dealt with statistical models. Finally, linguistic feedback is given to the operator using a fuzzy expert system. The method is illustrated by an example of a simple but ill-structured work strategy optimization problem associated with machine work.