Sistema dinámico para la detección de anomalías en entornos de Internet de las cosas (IoT)

Abstract

The Internet of Things (IoT), which refers to the connection of common objects to the internet, has produced numerous technological advances in domestic and industrial areas, among other sectors. These devices, equipped with sensors and actuators, are revolutionizing areas such as healthcare with telemedicine and the creation of smart cities. The leading companies have identified and invested in this potential, boosting its relevance. However, with the proliferation of connected devices come significant security challenges. IoT systems, with their variety of protocols and limited computational capacity, present unique vulnerabilities. Given the importance of cyber security for adoption and trust in IoT technology, IDS systems are a way to implement security by detecting potentially malicious traffic, without impacting existing systems. In addition, the incorporation of artificial intelligence techniques, such as machine learning, is becoming increasingly relevant to improve the efficiency of these systems. The research of this thesis aims, among other objectives, to gain an in-depth understanding of IoT security protocols and issues, to learn about current solutions, to develop a system for obtaining and analyzing traffic datasets. These datasets are of great importance because IDSs, particularly those based on anomaly detection, require relevant data oriented to the specific vulnerabilities of IoT systems. Also, as part of the results of this thesis, an IDS system is designed using machine learning models trained with the obtained datasets. This IDS will be evaluated in real environments to test the detection models in real IoT environments with specific threats.