Collaborative Radio Frequency Fingerprint Identification Using Dual-Channel Parallel CNN

The majority of current research in radio frequency (RF) fingerprint identification is conducted based on deep learning. In many existing RF fingerprint identification systems, a single receiver is typically utilized to collect data and perform RF sensing independently. This design can streamline the hardware structure of the system and reduce energy consumption and cost, but concurrently, it can also constrain the performance of the RF fingerprint identification system in terms of accuracy and robustness. Consequently, we propose a collaborative RF fing-erprint identification method based on a dual-channel parallel convolutional neural network (CNN), which employs dual receivers to collect data collaboratively and extracts RF fing-erprint features through a dual-channel parallel convolutional neural network to finalize RF fingerprint identification. The experimental results demonstrate that the RF fingerprint identification scheme proposed in this paper exhibits superior identification accuracy with the same amount of data compared to the RF fingerprint identification scheme with single-receiver independent identification.