Publicly Verify While Hiding Data: Privacy-Aware Streaming Truth Discovery with Public Verifiability in Blockchain-Enhanced Crowdsensing
Ruikai Zheng, Haiqin Wu, Boris Düdder · 2024
In crowdsensing, truth discovery (TD) has been extensively applied to resolve the data conflicts among publicly recruited workers and provide more reliable task data truths for task requesters. Data privacy and computation integrity are two major security concerns in TD due to the inherent untrustwor-thiness of the crowdsensing platform. Previous resolutions mostly focus on the design of privacy-preserving TD schemes while those leveraging blockchain to ensure TD integrity unfortunately sacrifice privacy or endure expensive on-chain overhead. In the presence of any misbehaviors, making TD integrity publicly and efficiently verifiable without compromising data privacy is imperative. This paper proposes p2STD, a Prlvacy-preservlng and Publicly verifiable scheme for generic Streaming TD (STD) in blockchain-enhanced crowdsensing. Unlike traditional TD in centralized crowdsensing, P2STD works in a decentralized architecture empowered by blockchain and fog, in which STD is jointly performed by fog nodes while the proof information is anchored to the blockchain for public verification. We take streaming confident-aware TD (S-CATD), the latest STD, as an instance and design secure aggregation protocols based on verifiable additive homomorphic secret sharing, concealing both task data and truth and generating publicly verifiable information. p2 STD is not specific to S-CATD and can be adapted to other iterative STD algorithms. Security analysis proves the desired privacy and public verifiability achieved in p2 STD. Experimental evaluations demonstrate that P2STD has high accuracy, lower computational and communication overhead, offering public verifiability with additional subtle gas cost than others.