Stacked LSTM-Based Radio Frequency Fingerprint Identification With Blind Equalization

作者：Juncheng Pan, Ying-Ke Lei, Caiyi Lou, Fei Teng · 发表于：2024 4th International Conference on Neural Networks, Information and Communication (NNICE) · 年份：2024 · DOI：10.1109/nnice61279.2024.10498843 · 被引用次数：4

As more and more wireless devices are widely connected to the Internet of Things(IoT), the authentication and identification technology of devices has become a key instrument in ensuring the security of IoT. The utilization of Radio Frequency Fingeprint (RFF)-based device identification technology can effectively achieve physical layer authentication of IoT devices. In this paper, a blind equalization (BE) based RFF extraction method is proposed, which can be performed during the process of receiving the signal without additional processing. A three-layer stacked long short-term memory (SLSTM) network is designed for employing BE-based RFF to identify different devices. The BE-SLSTM proposed in this paper achieves an identificaiton accuracy of 96.43% under a signal-to-noise ratio (SNR) level of 10dB and 99.36% under 30dB, as evidenced by the experimental results from 10 radio stations.