부산대학교 도서관

Whereas the deep learning (DL) methods were already widely used in state-of-the-art Computer Vision (CV) as well as Natural Language Processing (NLP) implementations, it has only been in recent times that these innovations have begun to grow up a bit adequately in applications involving wireless communications, a field known as Radio Frequency Machine Learning (RFML). Particularly noteworthy is new studies that has demonstrated that deep learning (DL) is an enabler innovation for Cognitive Radio (CR) applications, as well as a valuable tool for enhancing the performance of skillfully specified techniques for frequency recognition applications including such signal identification, prediction, or categorization. Since there is an availability of comparable raw Radio Frequency (RF) information available to enable learning and assessment, almost no previous information about the target spectrum environment is necessary. This is a primary motivation for the use of RFML. Apart from the basic necessity for adequate information, there are many other important aspects to taken into perspective before using RFML systems in genuine wireless communication applications. These concerns include trustworthiness, security, and technical specifications. The sharing and disseminated nature of wireless medium makes wireless systems susceptible to a variety of threats such as interference and surveillance. Machine learning (ML) offers an autonomous technique of learning from and adapting to wireless communication features that are difficult to obtain using custom characteristics & algorithms, which may be used to assist both attacking and defensive tactics. Based on this research ways which support the application of machine learning in enhancing the security.