Edge-Assisted Public Key Homomorphic Encryption for Preserving Privacy in Mobile Crowdsensing

Abstract

Mobile crowdsensing (MCS) is becoming an increasingly important topic due to rapid proliferation of mobile apps where participants’ anonymity is a pivotal requirement with direct impacts on their safety and well being. There are two main challenges in crypto-based privacy-preserving aggregation in MCS, namely, participants joining or leaving the crowd randomly at will, and adversaries injecting fake data. The conventional approach for preserving privacy is to provide blanket anonymity to all, including adversaries, which enables the latter to cause harm without being identified. In addition, with the proliferation of edge servers, there is a need to develop edge-assisted MCS, which would be more efficient in terms of less back-haul traffic and less delay as compared to cloud-assisted-only MCS. In this paper, we present an efficient edge-assisted MCS scheme which preserves the participants’ privacy and anonymity, protects the service against adversaries, and can be used to verify that aggregation is free of anomalies. Our scheme is transparent to the join-and-leave problem; and its computational cost and communication overhead are small and fixed, i.e., it is insensitive to crowd count. We utilize group signature for source authentication to identify and block adversaries that cause harm while providing anonymity to ordinary participants.