Improved Swarm Learning with Differential Privacy for Radio Frequency Fingerprinting
Lei Zhang, Lei Feng, Yue Liu, Fanqin Zhou, Boyu Liu, Zheng Jia, Yanru Wang, Hui Liu, Wenjie Ma · 2023
As the multimedia technology evolves, the security of multimedia services becomes more and more important. Among the multimedia services components, the most important is the security of multimedia Internet of Things (IoT) device. Radio Frequency Fingerprinting (RFF) is a physical layer authentication factor with good anti-counterfeiting capability, which can be used for identity authentication to ensure the security of IoT devices. Since radio frequency fingerprinting is a crucial privacy data for IoT devices, this study utilized differential privacy to improve swarm learning technology and designs a malicious device evaluation method to enhance swarm learning’s privacy protection capability. Our goal is to improve the privacy protection capability of the swarm learning architecture so that IoT devices can combat the danger of RFF data leakage at the cost of a minimal amount of model precision, so as to lay a foundation for the security of multimedia services.