Identifying Internet of Things Devices through Unique Digital Signa-tures and Advanced Machine Learning Techniques
Taiwo Abdulahi Akintayo, Richards Obada Okiemute, Moyosore Celestina Owoeye, Oluwaseyi Sulaimon Balogun, Chadi Paul, Madumere Chiamaka Queenet, Ruth Onyekachi Okereke, Adedokun Seyi Adediran, Christian Chukwuemeka Nzeanorue, Egenuka Rhoda Ngozi, Chika Moses Madukwe · Path of Science · 2024
The rapid growth of the Internet of Things (IoT) has led to a surge in connected devices across various sectors, necessitating reliable device recognition techniques. Device fingerprinting, which involves analysing network behaviour, communication patterns, and hardware features, offers a solution. Our proposed method leverages machine learning algorithms to analyse and categorise device fingerprints, achieving exceptional accuracy in identifying diverse devices, including sensors, actuators, and intelligent appliances. Moreover, it effectively detects suspicious devices and has a low computational overhead, making it suitable for real-time deployment. Our model demonstrates its effectiveness through rigorous testing and validation on multiple IoT datasets. The benefits of device fingerprinting for IoT device identification include enhanced security, improved network management, and increased visibility into device behaviour, making it a valuable tool for IoT ecosystem management.