Ship Encounter Situation Recognition Based on Transformer

Intelligent ships represent a crucial trend in the development of the maritime industry and will become the predominant vessel form in the future. Intelligent ships must have efficient perception and situational analysis capabilities in complex navigation environments to achieve intelligent decision-making and safe navigation. Maritime traffic safety is a critical issue for the global shipping industry, and maritime situational awareness is essential for ensuring safe navigation in waterways. This paper addresses the problem of intelligent identification of potential navigational risks in ship navigation environments and proposes a Transformer-based approach for ship encounter situation recognition. This method utilizes Automatic Identification System (AIS) data to extract encounter features. Contextual Position Encoding and Coordinate Attention mechanisms are introduced into the model to capture spatial correlation and directional features, enhancing the accuracy of determining encounter situations. Through experimentation in the research area and comparison with other models, the results demonstrate that the proposed model achieves higher recognition accuracy in ship encounter situation identification. This research provides a beneficial new idea and technical approach for studying intelligent environmental perception and situational analysis for ships.