Research on Device Identity Authentication Method Based on RF Fingerprinting Genetic Portrait Analysis

With the rapid development of Internet of Things systems, their security issues have garnered significant attention. To ensure reliable communication between devices and improving the security of Internet of Things systems, the study of device identity authentication methods has become an urgent concern. This paper presents a novel approach for device identity authentication, leveraging Radio Frequency fingerprinting genetic portrait. This paper contributes in three primary aspects. Firstly, this paper extracts the Radio Frequency fingerprint gene portrait as the unique identity signature of the device. Secondly, to improve the accuracy of device identity authentication, this paper selects the stronger feature points in radio frequency signals and extracts the corresponding feature descriptors. Finally, this paper utilizes a Convolutional Neural Network to achieve identity authentication for legitimate devices. Experimental validation was conducted in two scenarios, yielding device recognition accuracy rates of 99.3% and 97.4%, respectively. In comparison to the most recent research, the device recognition rate has surpassed previous benchmarks by 10.0%.