Yasir Abdulzahra Flaiyh Alaabedi

Radio fingerprint aims to provide an authentication and authorization for wireless devices by identification. It improves the security of wireless network at the physical layer which also enhance the time complexity by eliminating the need of authentication at higher layers. Artificial intelligence (AI) has proven its significant improvement in the radio fingerprinting, especially a multidimensional mapping feature of AI enhances the accuracy of radio fingerprint. However, the wireless network channel reduces the accuracy of radio fingerprinting algorithm. Hence, it is important to study the data classification of radio fingerprinting with lager amount of dataset collected from various wireless channels and quantitative analysis of effect of wireless network on radio fingerprint. In this research, 8 terabytes of data are collected from various wireless devices having similar radio frequency circuit. This dataset is used to analyze the effect of wireless network on radio authentication. Experimental results show the effect of wireless network channel on the classification accuracy of artificial intelligence algorithm which varies from 98 to 8% for the collected dataset. It is also observed that balancing I/Q data increases fingerprinting accuracy.