Differential Privacy in Cognitive Radio Networks: A Comprehensive Survey

Abstract

Integrating cognitive radio (CR) with traditional wireless networks is helping solve the problem of spectrum scarcity in an efficient manner. The opportunistic and dynamic spectrum access features of CR provide the functionality to its unlicensed users to utilize the underutilized spectrum at the time of need because CR nodes can sense vacant bands of spectrum and can also access them to carry out communication. Various capabilities of CR nodes depend upon efficient and continuous reporting of data with each other and centralized base stations, which in turn can cause leakage in privacy. Experimental studies have shown that the privacy of CR users can be compromised easily during the cognition cycle, because they are knowingly or unknowingly sharing various personally identifiable information (PII), such as location, device ID, signal status, etc. In order to preserve this privacy leakage, various privacy preserving strategies have been developed by researchers, and according to us differential privacy is the most significant among them. In this article, we provide a thorough survey on how differential privacy can play an active role in preserving privacy of cognitive radio networks (CRN). Firstly, we provide a thorough comparison of our work with other similar studies to show its novelty and contribution, and afterwards, we provide a thorough analysis from the perspective of various CR scenarios which can cause privacy leakage. After that, we carry out an in-depth assessment from the perspective of integration of differential privacy at different levels of CRN. Then, we discuss various parameters which should be considered while integrating differential privacy in CRN alongside providing a comprehensive discussion about all integrations of differential privacy carried out till date. Finally, we provide discussion about prospective applications, challenges, and future research directions. The discussion about integration of differential privacy in different CR scenarios indicates that differential privacy is one of the most viable mechanisms to preserve privacy of CRN in modern day scenarios. From the discussion in the article, it is evident that the proposed integration of differential privacy can pave the way for futuristic CRN in which CR users will be able to share information during the cognition cycle without the risk of losing their private information.