Aiding the analysis of human actions in large-scale systems: an intelligent interface approach

Analysis of human behavior, going as far as errors, in large human-machine systems such as power plants and aircraft cockpits, is recognized as one of the most essential tasks for enhancing system safety. As the cognitive aspects of human-machine interaction are increasingly emphasized, the task of description and analysis of human activities becomes more complicated and demanding. This paper presents a support system which implements some original concepts of intelligent interface to support the analytic description of human cognitive actions and the human-machine interaction that surrounds them. The aid enhances the coherence and the resolution of the incident analysis so that necessary counter measures to overcome the found errors can be clearly derived. It employs both model-based and rule-based reasoning based on an operator model, a rule base of operational constraints, and an extensive knowledge base that represents the system configurations. The interface was designed to be transparent and non-intrusive considering the cognitive process of the user. The system was developed for incident analysis in nuclear power plants in Korea, but the approach is gneral enough that it may be applied to the analysis of human-machine interactive behavior of any large-scale systems.