Internal Representation, Internal Model, Human Performance Model and Mental Workload

Abstract

To supervise correctly a complex technological plant, the human supervisor must be very familiar with the plant, that is he must possess an INTERNAL REPRESENTATION, IR, of the plant, the tasks, and the statistics of disturbances, if he or she is to act in an optimal way. The researcher's understanding of the supervisor's IR is an INTERNAL MODEL, IM, which can be included in an overall model of the supervisor's behavior, called HUMAN PERFORMANCE MODEL, HPM. Over the last decades control, AI, verbal and psychological models have been proposed. The existence of an IM is a necessary but non-sufficient condition for building HPM's. Models have been developed at the skill-based and rule-based level, but not at the knowledge-based level, yet it is just the human supervisor's creativity that motivates the need to maintain the human supervisor's job; it is his knowledge-based behavior that cannot be modeled, and hence it is of great interest for scientists, engineers and product managers.Mental load has also received much attention recently, but almost none has been given to the combination of human performance and mental load during human supervisory tasks. The relation of mental load, performance and IR has not been considered in depth. Very few results are published and no field evaluations have been reported.This paper reviews the progress on human operator modeling in terms of performance and mental load; it will elucidate the importance of the IR in these, and finally the practical value of human modeling will be discussed.