HEMETS - Human Error Modeling for Error Tolerant Systems

D.N. American proposes the development of a Human Error Modeling tool for Error Tolerant Systems (HEMETS) which will simulate human performance on complex systems in order to provide system designers with information on what aspects of their design are likely to lead to human error. Furthermore, HEMETS will provide recommendations to the designers so they can improve their design in order to eliminate, or at least reduce the effects of, such errors. HEMETS builds upon a well established computational cognitive architecture, (ACT-R) and, utilizing an analysis of research in human errors, establishes a cognitive architecture for human error modeling - the Human Error Modeling Architecture (HEMA). In order to enable HEMETS to assess a wide variety of designs HEMETS provides a HEMA System Interface (HEMA SI). This will support communication between HEMA and a simulated system user interface at the level of the visual display, allowing HEMETS to function with most prototype user interfaces for proposed systems. The recommendations for the system designers will be provided in the form of an automated Model of Human Error Prediction (MHEP). The MHEP will be based on an analysis of the needs of system designers and apply the team’s expertise in cognition, perception and human factors to provide feedback in a form conducive to improving a system design. The MHEP will be of a form comparable to the explanatory subsystems of expert systems.