An Experimental Study on Implementing Real-time Estimator of Human Cognitive Process into Eye-Sensing Head Mounted Display for Realizing Adaptive CAI for Teaching Plant Diagnosis Knowledge

摘要

Computer and information technologies have been advancing so rapidly, and by introducing automation, the engineering systems these days become more "useful" than before. However, it is not necessarily the situation that present systems have become "usable" for users. In order to enhance "usability" of the system, it is very important for man-machine interfaces to provide a friendly communication function which offers adaptive advice for individual users. The authors have proposed a concept of adaptive interface which has the adaptation functions by considering various human factors (Takahashi et al. 1993). This will be made by on-line measurements of human information activities to judge users' internal states such as thinking process, mental workload, and then by adaptive instruction to users in accordance with the estimated internal states. The authors have aimed at applying such adaptive interface concept for constructing a computer-assisted instruction (CAI) system for the training of knowledge and skill necessary for diagnosing nuclear power plant anomalies. In order to provide the CAI system with appropriate instructions for each trainee, the system should understand the trainees' personal thinking process promptly through his interactive dialogue with the interface. In this respect, there are two types of information which are necessary for tracing trainees' thinking process. One is concerned with "which plant parameter is he/she looking at?", while the other is "how does he/she judge on the parameter he/she is looking at now?". In this study, the authors have tried to acquire the first type of information from trainees' eye gaze points, while the second one, from verbal information.