RF Fingerprinting Techniques for Passive Identification of IoT Devices
Keywords:
RF fingerprinting, IoT device identification, wireless security, passive sensing, signal classification, machine learning.Abstract
At the device level, identification is an important issue in protecting wireless Internet of Things (IoT) infrastructures. The study presents an effective radio-frequency (RF) fingerprinting model of passively identifying IoT devices. Using inherent hardware flaws like I/Q imbalance, transient response and spectral distortion, the proposed system uses these flaws to create distinct RF signatures. A convolutional neural network (CNN) classifier is used to process these signatures and the recognition accuracy is 97 % of 50 types of devices. The system does not require any cryptographic keys or active handshake protocols, and provides a lightweight and scalable authentication layer to low-power IoT systems. The experimental findings support the robustness of the method when applied in the dynamic conditions of the channel, which proves that the method can support the development of device-level trust in the situation of large-scale networks.
