Analyzing adversary capability for subset coding based privacy protection in participatory sensing

Abstract

With the growth of network communication across the world, Participatory Sensing Sys-tem (PSS) is being acknowledged as an emerging technology given its potential for a wide variety of sensing applications. As the participation of users is the key to success of PSS applications, protecting the privacy of their shared information is the inevitable challenge to deploy PSS in practical scenarios. Many researches on this eld aim to solve this chal-lenge of protecting privacy as well as achieving exact data recoverability. Anonymization based approaches are so far the most acceptable techniques to provide both privacy and data quality in terms of data recoverability. Among them, subset-coding based privacy-protecting anonymization achieves data quality with acceptable computational complexity. However, this technique also su ers from the privacy risk for trusting additional anonymiza-tion server. Moreover, the adversary risk considering di erent relevant parameters is not analyzed in the anonymization schemes proposed so far. This research study aims to provide a privacy-protective solution without the help of additional anonymizaton server and to an-alyze its privacy risk against adversary attempts. The proposed scheme achieves acceptably close performance to those using dedicated third-party sever as established with theoretical analysis and comprehensive simulation. Our proposed scheme does not degrade in terms of computational complexity, user participation requirement or privacy protection in spite of removing additional anonymization server as established with theoretical analysis and comprehensive simulation.