A Novel Framework for Open-Set Authentication of Internet of Things Using Limited Devices

Area of Science:

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

Background:

• The Internet of Things (IoT) offers transformative potential across various sectors.
• The open architecture of IoT systems presents significant cybersecurity vulnerabilities, including identity spoofing.
• Physical layer authentication offers a robust defense against identity spoofing by leveraging unique signal characteristics.

Purpose of the Study:

• To develop a deep learning-based framework for open-set authentication of IoT devices.
• To improve the accuracy and reliability of IoT device identification and unauthorized access prevention.

Main Methods:

• Implementation of additive angular margin softmax (AAMSoftmax) to improve feature discriminability.
• Utilization of a modified OpenMAX classifier for adaptive identification of authorized and unauthorized devices.
• Testing with both simulated and real-world Automatic Dependent Surveillance-Broadcast (ADS-B) data.

Main Results:

• The proposed framework demonstrated superior performance in authenticating IoT devices.
• Effectiveness was particularly notable in scenarios with a limited number of training devices.
• The system successfully distinguished authorized devices from unauthorized ones in open-set conditions.

Conclusions:

• The deep learning framework provides an effective solution for IoT device authentication and security.
• The approach enhances the security posture of IoT ecosystems against sophisticated cyber threats like spoofing.
• This method offers a promising direction for securing the growing landscape of connected devices.