Secure Collaborative Sensing for Crowd Sourcing Spectrum Data in White Space Networks

Collaborative Sensing is an important enabling technique for realizing opportunistic spectrum access in white space (cognitive radio) networks. We consider the security ramifications of crowdsourcing of spectrum sensing in presence of malicious users that report false measurements. We propose viewing the area of interest as a grid of square cells and using it to identify and disregard false measurements. The proposed mechanism is based on identifying outlier measurements inside each cell, as well as corroboration among neighboring cells in a hierarchical structure to identify cells with significant number of malicious nodes. We provide a framework for taking into consideration inherent uncertainties, such as loss due to distance and shadowing, to reduce the likelihood of inaccurate classification of legitimate measurements as outliers. We use simulations to evaluate the effectiveness of the proposed approach against attackers with varying degrees of sophistication. The results show that depending on the attacker-type and location parameters, in the worst case we can nullify the effect of up to 41% of attacker nodes in a particular region. This figure is as high as 100% for a large subset of scenarios.