As one of the key technologies for realizing a fully digitalized world, the Internet of Things (IoT) requires ubiquitous connections across both land and sea. However, due to lack of infrastructure such as optical fibers and base stations, maritime communications inevitably face a highly complex and heterogeneous environment, which greatly challenges the connection reliability and traffic steering efficiency for future service-oriented maritime IoT. With the recent burgeoning application of artificial intelligence (AI) in many fields, an AI-empowered autonomous network for maritime IoT is envisioned as a promising solution. However, AI typically involves training/learning processes, which require realistic data/environment in order to obtain valuable outcomes. To this end, this article proposes the parallel network, which can be regarded as the "digital twin" of the real network and is responsible for realizing four key functionalities: self-learning and optimizing, state inference and network cognition, event prediction and anomaly detection, and knowledge database and snapshots. We then explain how various AI methods can facilitate the operation of the parallel- network-driven maritime network. A case study is provided to demonstrate the idea. Research directions are also outlined for further studies.

中文翻译：

---

人工智能赋能的海事物联网：并行网络驱动的方法


物联网作为实现全数字化世界的关键技术之一，需要陆地和海洋的无处不在的连接。然而，由于光纤、基站等基础设施的缺乏，海上通信不可避免地面临着高度复杂和异构的环境，这对未来面向服务的海上物联网的连接可靠性和流量引导效率提出了极大的挑战。近年来，随着人工智能（AI）在许多领域的蓬勃发展，人工智能赋能的海事物联网自主网络被认为是一种有前景的解决方案。然而，人工智能通常涉及训练/学习过程，需要真实的数据/环境才能获得有价值的结果。为此，本文提出了并行网络，它可以被视为真实网络的“数字孪生”，负责实现自学习和优化、状态推理和网络认知、事件预测和异常等四个关键功能。检测、知识库和快照。然后，我们解释各种人工智能方法如何促进并行网络驱动的海事网络的运行。提供了一个案例研究来证明这个想法。还概述了进一步研究的研究方向。