Intelligent Maritime Collision Avoidance and Early Warning System for Navigation Safety

Abstract

This paper presents the design and implementation of an intelligent ship collision avoidance and early warning system aimed at enhancing maritime navigation safety. The system integrates data from multiple sources-including AIS, radar, GPS, and marine electronic charts-to achieve real-time monitoring, trajectory prediction, and collision risk assessment. An Attention-LSTM-based trajectory prediction model is developed to forecast vessel movements with high temporal accuracy, while visual target recognition and ship nameplate identification modules based on convolutional neural networks (CNNs) enable visual perception under complex maritime conditions. In addition, a Unity3D-based dynamic simulation platform is constructed to validate the collision avoidance strategies and visualize ship interactions in various scenarios. Experimental results indicate that the proposed system effectively predicts navigation trajectories, improves detection accuracy, and significantly reduces potential collision risks. The research provides a practical and extensible framework for intelligent maritime navigation systems and supports future applications in coastal management, marine traffic control, and autonomous ship operation.