Navigation Scenario Permutation Model for Training of Maritime Autonomous Surface Ship Remote Operators

The development of autonomous ships has begun. Artificial intelligence (AI) is expected to be partially responsible for navigation; nevertheless, the importance of human intervention is higher than ever. Human intervention in the control of an autonomous ship via the remote operator requires navigation proficiency. The education method for the remote operators that is presently considered is simulation training. However, the simulation training does not take long enough time for enabling trainees to develop their navigation proficiency equivalent to that of conventional ships navigators. In addition, the simulation training should contain various navigation scenarios to train the trainee properly. Therefore, this paper suggests the methods to generate the massive and practical navigation scenarios by extracting navigation elements' distribution from actual ship trajectory data and applying them to the permutation of navigation elements. The results demonstrated the advantages of the proposed methods by comparing the sample navigation scenario and an example of an impractical navigation scenario. In conclusion, it is expected that the massive generation of practical navigation scenarios using the proposed permutation model will positively affect the simulation training of the maritime autonomous surface ship remote operators.