It is essential to adapt the current traffic rules to meet the implementation of Maritime Autonomous Surface Ships (MASS) being developed worldwide. Current rules are designed with vessels operated by a person on board in mind. Therefore, it is expected to be difficult to apply these rules to an MASS. The present study attempted to analyze the behavior patterns of seafarers when they take evasive action and quantify the degree of ambiguity in their actions concretely. Analysis of the patterns of collision avoidance among active seafarers revealed no consistency in their behavior. It is argued that the influence of an individual seafarers’ experience and personality presents a significant influence on ship operation.

For autonomous underwater vehicles (AUVs) navigating underwater, where satellite positioning information is not available, procedures for identifying and correcting positions using acoustic positioning from the support ship are important. In this study, we propose a quick and accurate method to perform position error correction using the positions obtained from a support ship and dead-reckoning estimation. The dead-reckoning estimation was transmitted from an AUV during its pre-survey steady turn, which was performed near the sea bottom. The center position of the circular trajectory obtained by the steady turn was detected using the Hough transform, which was applied to each of the dead-reckoning positions of the AUV received by underwater acoustic telemetry and the positions obtained by underwater acoustic positioning conducted from the support ship. The position error was estimated from the distance between the centers of the two circular trajectories and sent to the AUV to correct its dead-reckoning position. Three dives were conducted during a sea trial carried out in Suruga Bay, Japan. We could significantly reduce the position error, resulting in a minimum error of 1.3 m. The proposed method has a sufficiently high degree of accuracy for practical purposes and low calculation costs for on-site applications.