Privacy-Preserving Crowd-Sensing for Dynamic Spectrum Access with Malicious Workers

Abstract

The rapid growth of IoT wireless devices has caused severe spectrum shortage. The Dynamic Spectrum Access (DSA) paradigm is extremely well-suited to alleviate such spectrum shortage. Contrary to legacy spectrum distribution methods, DSA allows for spectrum reuse whenever the primary licensees are idle and not actively utilizing the spectrum. DSA's database-driven model has been very successful and applicable towards TV white-space frequencies. However, other more fluid spectrum opportunities, such as the National Oceanic and Atmospheric Administration spectrum blocks, continue to be identified and categorized via either (i) solitary or (ii) multinodal cyclostationary and cooperative signal detection/classification methods. The later methods improve significantly the rate of signal detection and classification yet they have severe drawbacks when it comes to preserving the location privacy of users. Moreover, both methods are negatively affected by large delays, partially due to the quiet periods they need to perform while identifying/classifying incumbents. To this end, we extend our previous work by proposing a novel privacy-preserving scheme based on the fusion of DSA and crowd-sensing paradigm. Our scheme provides location privacy for the crowd-sensing users through homomorphic cryptographic constructions. Furthermore, our scheme mitigates fraudulent sensing report attacks by providing robust fraud prevention based on homomorphic cryptographic constructions and reputation algorithms. We provide substantial experiments on real-life datasets which shows that our proposed protocol provides realistic efficiency and security.