Navigating Patterns Analysis for Onboard Guidance Support in Crossing Collision-Avoidance Operations

Abstract

A proper interpretation and classification of navigators’ operational behaviors is crucial to the design of onboard decision-support systems. This research work dives into the study of navigators’ navigating patterns (NPs) in a maritime collision-avoidance (CA) traffic situation. Three NPs, specifically conservative, moderate, and aggressive modes, are identified with respect to a collision risk assessment (CRA) by interpreting data collected from the GPS and automatic identification systems. The CRA is realized following the collision risk modeling concept of the closest point of approach. Then, a human-centered onboard guidance-support system is developed according to the patterns identified to help navigators make decisions. This proposed approach is implemented in the scenario of sailing across a narrow strait, where human intelligence remains necessary in the foreseeable future. The research experiment was conducted on Kongsberg maritime simulators. Thirty-six rounds of sailing data containing 108 CA subtasks were collected and analyzed to classify NPs. Afterward, a guidance-support system was designed based on the patterns’ demonstration. An additional experiment to test the developed system in the same scenario was organized on the same simulator. The results show that the system can considerably improve the navigator’s navigation management ability in CA operations. Our approach combines data analysis and risk modeling with authentic human-operated navigating data and traffic information, which makes it distinct from traditional intuitive and cognitive maritime traffic modeling. It is the first one that defines NPs and puts them into potential industrial application pragmatically.