Paper Title
AIS–IoT Fusion-Based Machine Learning–Driven Smart Maritime Vessel Monitoring and Collision Avoidance System

Abstract
The transportation by sea has experienced a remarkable rise in the number of vessels in ports, coastal areas and narrow waterways and this has greatly amplified the chances of vessels colliding and having a navigational accident. Although the Automatic Identification System (AIS) allows the automatic identification of vessels and the real-time tracking of them, the system alone does not offer predictive intelligence and proactive collision avoidance. The current paper introduces an AIS-IoT fusion-oriented, machine learning-powered smart maritime vessel monitoring and collision avoidance system that aims at increasing the real-time situational awareness and providing proactive navational decision-making. The suggested model incorporates AIS path information and IoT-based environment, such as wind, and visibility, and evaluates the vessel movement aspects, such as relative distance, speed, heading, and collision convergence, to determine the risk of collision levels. The concept of lightweight learning-inspired models is implemented on a Raspberry Pibased edge computing platform to allow processing low-latency and ensured functionality of resource-constrained hardware. Simulated and real AIS data assessment in a variety of vessel encounter scenarios, such as head-on, crossing, and overtaking, prove the correct estimation of real-time risks of collision and the timely provision of alerts, which is why the suggested method can be considered effective in terms of smart monitoring of traffic in the sea and the improvement of the navigational security. Keywords - AIS, IoT, Maritime Collision Avoidance, Machine Learning, Edge Computing, Vessel Monitoring