A Short Survey on Embedded AI with Edge Computing for Maritime Applications

Embedded AI systems, when integrated with Edge Computing, enable real-time processing and data analytics on-board maritime operations, effectively addressing power and computation limitations. This paper reviews such integrated systems, explores their benefits, and examines their implications for maritime applications. We discuss relevant technologies and applications, focusing specifically on AI systems deployed at sea. The implementation of Convolutional Neural Networks, particularly YOLO, is analysed across maritime applications ranging from object detection to autonomous vehicle operations. We also explore how Internet of Things (IoT) devices and sensors enhance the performance of embedded AI systems in maritime environments. Our analysis identifies four main areas: maritime object detection, marine ecosystem monitoring, autonomous vehicle operations, and maritime search and rescue missions. Results indicate that embedded AI systems significantly improve operational efficiency and enable new capabilities in maritime applications, particularly in autonomous vehicle operations and real-time monitoring systems. However, addressing resource constraints and improving system reliability in harsh maritime conditions remains a challenge for future research.