Section: New Results

Towards privacy-sensitive mobile crowdsourcing

We obtained new results in the domain of data privacy for crowdsourced data.

We proposed an anonymous data collection library for mobile apps, a software library that improves the user's privacy without compromising the overall quality of the crowdsourced dataset. In particular, we proposed a decentralized approach, named FOUGERE, to convey data samples from user devices using peer-to-peer (P2P) communications to third-party servers, thus introducing an a priori data anonymization process that is resilient to location-based attacks. To validate the approach, we proposed a testing framework to test this P2P communication library, named PeerFleet. Beyond the identification of P2P-related errors, PeerFleet also helps to tune the discovery protocol settings to optimize the deployment of P2P apps. We validated FOUGERE using 500 emulated devices that replay a mobility dataset and use FOUGERE to collect location data. We evaluated the overhead, the privacy and the utility of FOUGERE. We showed that FOUGERE defeats the state-of-the-art location-based privacy attacks with little impact on the quality of the collected data [38], [5].

These results have been obtained in the context of the PhD thesis of Lakhdar Meftah [13] defended in December 2019.