Editorial: Machine learning for resource management in industrial Internet of Things

摘要

Recent advances in the Industrial Internet of Things (IIoT) field have seen significant growth in the last couple of years. IoT is revolutionizing manufacturing, transportation, oil \& gas, and logistics sectors. However, developing IIoT applications poses several challenges, including the limited computational, memory, and energy resources of IoT devices. These devices generate a large amount of data at the network edge, making cloud-based processing impractical due to bandwidth constraints, latency, and security risks. Edge computing, which brings data processing closer to the source, offers a viable solution to these challenges. Despite its promise, edge computing faces significant hurdles. One of the primary challenges lies in the diversity of sensor types deployed across different environments, which adds complexity to the system's architecture.Furthermore, the large-scale deployment of edge devices and the inherent resource constraints of these devices complicate the task of optimizing performance. Machine learning has emerged as a powerful tool to address these issues, particularly in domains like robotics and natural language processing, where it helps optimize task allocation, improve decision-making processes, and enhance the overall efficiency of edge systems. This Research Topic features four articles that explore diverse and cutting-edge applications within the IIoT, including resource management and enhancing security in IoT systems. The first article, \textit{"An enhanced whale optimization algorithm for task scheduling in edge computing environments"} by \href{https://doi.org/10.3389/fdata.2024.1422546} {Han et al.}, focuses on addressing the challenges in real-time execution due to limited resources in edge computing environments. The authors proposed an enhanced whale optimization algorithm incorporating a multi-objective model considering CPU, memory, time, and resource utilization for optimizing task scheduling in edge computing. By leveraging chaotic mapping and a nonlinear convergence factor, the algorithm balances local and global search, significantly reducing costs (by 29.22\%), completion time (by 17.04\%), and improving resource utilization (by 9.5\%). This work significantly addresses the increasing demand for real-time processing capabilities in resource-constrained edge environments.The second contribution is the comprehensive review titled \textit {"Unveiling the core of IoT: comprehensive review on data security challenges and mitigation strategies"} by \href{https://doi.org/10.3389/fcomp.2024.1420680}{Kaur et al.}, which examines the security challenges posed by the increasing complexity of IoT environments. The authors identified key security threats, including spoofing, distributed denial of service, and man-in-the-middle attacks. This paper reviews various mitigation strategies such as machine learning, deep learning, lightweight encryption, intrusion detection systems, and advanced security protocols. The evaluation of IoT technology, the accompanying security progress, and the need for continued development are discussed. The paper also identifies IoT's application areas, such as healthcare, smart cities, smart homes, and industrial IoT, highlighting specific security challenges each faces. This review provides valuable insights into current vulnerabilities and presents strategies that could significantly enhance the resilience and security of IIoT systems. In the third article, \textit {"'Below 58 BPM,' involving real-time monitoring and self-medication practices in music performance through IoT technology"} by \href{https://doi.org/10.3389/fcomp.2024.1187933}{Merendino et al.}, the authors explored the development of an Internet of Musical Things system designed to assist an opera singer with a carotid aneurysm during performances. This system monitors the singer's heart rate in real-time and promotes self-healing by providing non-int