Model-Based Prediction of Human Performance with Respect to Monitoring and Failure Detection

Abstract

Human operator models for monitoring and failure detection are outlined. Corresponding experimental paradigms and extensive validation studies including eye movement results are explained. The state-space oriented models proceed with the information structure of the optimal control model (OCM) and consider single-observation as well as sequential decisions. The broad coverage of the models opens analysis and design applications in the area of supervisory control. The paradigm refers to multiple-process situations withoptional dynamics, couplings and event characteristics and, thus, increases the practical utility of models. An outlook is given on design theory and respective methodologies for man-machine systems.