Radar Detection Support System Using Artificial Intelligence for Telexistence

Abstract

Marine navigation officers perform lookouts, which include visual observation and the use of radar to keep track of the surrounding conditions. Early detection of other ships and obstacles is crucial for ensuring safe navigation, particularly in congested areas. In previous studies, remote maneuvering was achieved using the same communication and steering as ordinary maneuvering. However, when performing a maneuver to avoid another ship, it is necessary to share information with other ships remotely. Therefore, the purpose of this study was to obtain information on other ships of all sizes, by analyzing radar images. Artificial intelligence (AI) was used to predict and enhance radar echoes for small vessels. The tracking performance was compared between normal and AI-processed images using actual radar images obtained in Tokyo Bay. The results showed that the tracking performance of the AI system was better than that of the normal radar images. In addition, the AI-processed images could capture and track a small vessel at an early stage, before the navigation officer detected it.