RF Signal Feature Detection and Identification Method Unaffected by Complex Environments Based on I/Q Data

Abstract

With the advancement of the Internet of Things (IoT) and the proliferation of frequency-using devices, incidents such as the misuse of electromagnetic spectrum resources, the establishment of unauthorized radio stations to transmit illegal information, and the occupation of primary user spectrum resources by unknown signal sources have occurred with increasing frequency. These incidents pose serious risks to society, resulting in adverse effects and significant harm. Radio frequency (RF) fingerprinting technology offers a new approach to wireless network security. RF fingerprints are based on the physical level of the device and cannot be easily changed. Even devices produced from the same batch have subtle differences. This paper proposes a unique identity feature detection and identification method of RF signal unaffected by complex environment. Based on distributed electromagnetic spectrum system, I/Q data of target signal is acquired. The I/Q distance-received signal strength (RSS) characteristic curve is obtained by processing and further processed to obtain the characteristic parameters. After the features of the target radio device are extracted, they are stored in the RF fingerprint database. Feature identification is achieved by comparing the newly extracted features with those already present in the database. Finally, the target signal is marked according to the feature. The experiments verify that the method has good environmental invariance and discrimination. This allows for direct association between a detected signal and its stored fingerprint.