IoT Devices Identification Based on Machine Learning

Abstract

Intelligent network has been becoming one of the main development directions of 5G. A massive number of Internet of Things (IoT) devices are flexibly connected to 5G network, which brings significant benefits for various practical services and applications. However, the terminal diversity makes the efficient management of terminals extremely challenging. It is difficult for operators to be fully aware of their IoT devices, let alone to provide personalized quality assurance. Therefore, accurate classification of IoT devices based on the data in the realms of the core network is of great significance. To address the problem, Machine Learning (ML) or Artificial Intelligence (AI) is applied. In this paper, we first propose intelligent network architecture. Then, we propose a based-ML approach to classify mobile IoT devices by using IoT devices' position data. We evaluate and compare the performance of five machine learning algorithms in classifying 4 different mobile devices. We analyze the impact of different feature sets. Our preliminary results show that IoT devices can be identified and classified by using a lightweight feature set based on ML.