Application of Radar Measurements in Neural-Dynamic and Game-Based Safe Ship Controls in Multi-Object Passing Situations.

Radar sensors, which enable the identification of the navigational situation and a preliminary assessment of collision risk, form the basis for safe ship control. This paper presents the author's quantitative methods for assessing collision risk in the context of multi-ship navigation, and then synthesizes safe control methods. The author proposes four neural domain variants and three mathematical models of collision risk. Proprietary methods are presented: neural-dynamic control (NDC) and game-based control (GC). An experimental comparison of these control methods is conducted using data from real-world navigation scenarios. Therefore, under favorable traffic conditions, the NDC method proved to be the most effective, while the GC method facilitated effective cooperative and non-cooperative management in situations of restricted ship traffic. The safe control methods proposed in this study will contribute to increased navigation safety, particularly in situations of high ship traffic density and challenging environmental conditions.