A New Energy-Aware Technique for Designing Resource Management System in the 5G-Enabled Internet of Things Based on Kohonen's Self-Organizing Neural Network

• 出版年: TRANSACTIONS ON EMERGING TELECOMMUNICATIONS TECHNOLOGIES
• 主要な著者: Zou, Yan; Cao, Q.; Nazif, Habibeh
• フォーマット: 論文
• 言語: English
• 出版事項: WILEY 2025
• 主題: Telecommunications
• オンライン･アクセス: https://www-webofscience-com.uitm.idm.oclc.org/wos/woscc/full-record/WOS:001445115200001

The Internet of Things (IoT) has accelerated the connectivity between physical objects and the Internet. It has become common to integrate IoT devices into our lifestyles, considering the fact that they make traditional devices to be more intelligent and self-sufficient. The usage of 5G-enabled IoT can be one such improvement, as it integrates multiple devices and allows for effective interaction and data sharing. However, with the growing extreme increase in the number of devices being connected, resource utilization efficiency has emerged as one major challenge. Comparing the existing resource management strategies with the current environment brought by even more complex IoT, the former have consistently failed, leading to the wastage of too much energy. Resource allocation and efficient utilization in IoTs encompass processing power, bandwidth, and energy for the appropriate and effective functioning of devices and networks. The conventional designs are inherently inefficient in that they cannot match with the pace and nature of IoT data structures, hence making it difficult to achieve any meaningful performance, and resources are also wasted in the process; thus, there exists the necessity for energy-efficient approaches that are adaptable to dynamic workloads. In consideration of the aforementioned factors, this paper proposes an entirely new approach employing a Kohonen neural network to address the issue of resource allocation with a focus on energy efficiency. The first of these steps is the collection of data obtained from IoT devices and the processing of this data in order to detect the important features; the second step is the usage of the algorithm to produce a resource map indicating the spatial distribution of resources, and the final step is the real-time modification of the resource map by incoming data to promote appropriate resource allocation. The analysis shows that when using the method provided, energy, costs, and delays in the implementation of the process have improved.