The significant advancements in sensors and other resource-constrained devices, capable of collecting data and communicating wirelessly, are poised to revolutionize numerous industries through the Internet of Things (IoT). Sectors such as healthcare, energy, education, transportation, manufacturing, military, and agriculture stand to benefit. IoT is expected to play a crucial role in implementing both Industry 4.0 and its successor, Industry 5.0. IoT relies on data collected by sensors from various points, shared over wireless or wired networks, making it more vulnerable to attacks. Consequently, addressing privacy and security concerns is of paramount importance for the widespread adoption of IoT across industries. Recognizing the pivotal role of IoT security, recent years have witnessed a marked upswing in publications dedicated to leveraging Machine Learning techniques for intrusion detection within the IoT framework. This paper embarks on a comprehensive endeavor to classify and characterize the myriad of intrusion detection methodologies that have emerged through the fusion of Machine Learning and IoT security. Serving as a timely and insightful review, this survey is not only of immense value to seasoned researchers immersed in this dynamic field but also serves as an invaluable resource for newcomers eager to contribute to the enhancement of IoT security. This paper sets itself apart from existing surveys by placing particular emphasis on recent advancements in machine learning-based intrusion detection across various IoT domains. Unlike previous surveys, it comprehensively explores papers published within the past five years, encompassing a wide range of dimensions within this field. These dimensions include, but are not limited to, medical IoT, agricultural IoT, industrial IoT, Fog/Edge IoT, Intelligent Transportation Systems, Smart Home Networks, and more.

中文翻译：

---

最近在机器学习驱动的物联网入侵检测系统方面的努力


传感器和其他资源有限的设备取得了重大进步，能够收集数据并进行无线通信，有望通过物联网 (IoT) 彻底改变众多行业。医疗保健、能源、教育、交通、制造、军事和农业等行业将从中受益。物联网预计将在实施工业 4.0 及其继任者工业 5.0 中发挥至关重要的作用。物联网依赖于传感器从各个点收集的数据，并通过无线或有线网络共享，这使其更容易受到攻击。因此，解决隐私和安全问题对于物联网在各行业的广泛采用至关重要。认识到物联网安全的关键作用，近年来致力于利用机器学习技术在物联网框架内进行入侵检测的出版物显着增加。本文致力于对通过机器学习和物联网安全融合而出现的无数入侵检测方法进行分类和表征。作为一次及时而富有洞察力的回顾，这项调查不仅对沉浸在这一动态领域的经验丰富的研究人员具有巨大的价值，而且对于渴望为增强物联网安全做出贡献的新手来说也是宝贵的资源。本文与现有调查不同，它特别强调跨物联网领域基于机器学习的入侵检测的最新进展。与以往的调查不同，它全面探讨了过去五年内发表的论文，涵盖了该领域的广泛维度。 这些维度包括但不限于医疗物联网、农业物联网、工业物联网、雾/边缘物联网、智能交通系统、智能家居网络等。