НЕЙРОМЕРЕЖЕВА МОДЕЛЬ КОНТРОЛЮ АПАРАТНОЇ СУМІСНОСТІ КОМПОНЕНТІВ IOT-СИСТЕМИ

Autor: Artur V. Timenko, Vadym Shkarupylo, V. V. Smolii
Rok vydání: 2021
Předmět:
Zdroj: Visnyk of Zaporizhzhya National University Physical and Mathematical Sciences. :52-59
ISSN: 2518-1785
2413-6549
Popis: Nowadays, the Internet of Things paradigm is constantly becoming more and more widespread. Its implementation on a global scale requires the involvement of specialized technologies and tools. Among the directions to be worked out during that is to ensure the interoperability between the components of the corresponding systems. Given work is devoted to elaborate the specified direction. To this end, the following task is resolved: to develop the model of the Internet of Things system components interoperability control on a hardware plane – to provide an opportunity to detect the “potentially unsafe” components on time. In presented work, the constituent of the proposed complex approach to the Internet of Things system components interoperability is revealed. Named approach encompasses the aspects of both software and hardware planes of interoperability. Within the given paper, the emphasis is put on a hardware plane though. The model of hardware interoperability control between the components of the Internet of Things system has been proposed. The model is constructed on the basis of neural networks mathematical apparatus. Hardware interoperability between system components is approached from the standpoint of functional safety of a system as a whole. Devices based on ESP 8266 and ESP 8285 microcontrollers are considered as the components: in particular, due to being thoroughly widespread and accessible. Practical significance of obtained results: proposed model allows identify and eliminate threats to the functional safety of the system as a whole at a component level by way of assessing the actual state of the latter. Moreover, an instrument for predicting the number of such components during a given time interval has been proposed. Verification of the model proposed is based on solving the tasks of approximation and extrapolation. It has been demonstrated that implementation of the model allows to identify system components that need to be adjusted or replaced, i.e., violate the functional safety of the system as a whole. In turn, solving the extrapolation task makes it possible to predict the number of such components after a given time.
Databáze: OpenAIRE
Externí odkaz: