Combining advisory and credibility information as communication message from intelligent machine for efficient man-machine cooperation

Importance of credibility information supplied by multiple agents, each conducting a diagnostic task based on mutually dissimilar functioning principle, was examined in terms of diagnostic performance of a machine-intelligent system. The aim of the system is to provide dependable assistance to operators of a large-scale complex artifact in a complicated and inexperienced abnormal situation. The validity of the proposed design was evaluated through numerical experiments with a simulation code of a nuclear power plant. In spite of the complexity of the abnormal events, the diagnostic system demonstrated satisfactory performance by taking the credibility information into consideration