Development and Evaluation of a Novel IoT Testbed for Enhancing Security with Machine Learning-Based Threat Detection

Area of Science:

• Cybersecurity
• Machine Learning
• Internet of Things (IoT)

Background:

• The Internet of Things (IoT) presents significant security challenges due to device diversity and resource limitations.
• Traditional security solutions struggle in dynamic IoT environments, and realistic datasets for ML model training are scarce.
• Vulnerabilities in IoT devices expose them to cyber threats like data breaches and denial-of-service attacks, undermining ecosystem trust.

Purpose of the Study:

• To develop and implement a novel physical smart office/home testbed for evaluating ML algorithms in IoT security.
• To create a controlled environment for generating realistic IoT network traffic and simulating diverse cyber attack scenarios.
• To address the lack of high-quality datasets for training and validating ML-based IoT security solutions.

Main Methods:

• Development of a physical smart office/home testbed integrating various IoT devices (sensors, cameras, smart plugs).
• Simulation of realistic network traffic patterns and diverse attack scenarios (unauthorized access, network intrusions).
• Training and validation of ML models, including XGBoost and SVM, for anomaly detection using testbed-generated data.

Main Results:

• The XGBoost model achieved high balanced accuracy (up to 99.977%) on testbed data, comparable to benchmark datasets.
• The SVM model demonstrated improved performance on testbed data (up to 96.71%) compared to benchmark datasets.
• The testbed proved effective in generating realistic datasets and enabling robust security evaluations for ML algorithms.

Conclusions:

• The developed testbed is a valuable tool for advancing IoT security research by providing realistic evaluation capabilities.
• The findings highlight the potential of ML algorithms in detecting and mitigating IoT security vulnerabilities.
• This work contributes to building more resilient and adaptive security frameworks for critical IoT infrastructures.