Abstract
Most of major marine accidents were caused by human operator. Therefore, human behavior is greate importance to analysis for maritime safety. A simulation approach, for investigating critical risks under specific maneuvering conditions, has been recently utilized for its advantage to save cost and time. The present study constructs a cognitive model of a ship navigator based on the cognitive task analysis in which synchronized data from eye-movement recordings and verbal protocols was analyzed during a series of actual navigation sessions using a maritime simulator. The model describes a dynamic interaction between the navigator and the sates in maneuver and navigation environment for a course-tracking task in which a ship is navigated to follow a planned route as safety as possible in the narrow fairway. This model implemented on a Macintosh PC using a standard C code. The accuracy of the constructed model is examined by comparing the task performance simulated by the model with the observed performance by a human navigator on the maritime simulator. It could produce similar patterns of helm-command generations to those of the actual human navigator.
