Encrypted Data Aggregation in Mobile CrowdSensing based on Differential Privacy

The increasing sensing capabilities of mobile devices enable the collection of sensing-based data sets, by exploiting the active participation of the crowd. Often, it is not required to disclose the identity of the owners of the data, as the sensing information are analyzed only on an aggregated form. In this work we propose a privacy-preserving schema based on differential privacy which offers data integrity and fault tolerance properties. In our schema, data providers firstly add a noise component to the sensed data and, secondly, they encrypt and send the cryptogram to the aggregator. The data aggregator is in charge of only decrypting the cryptograms, by preserving the identify of the data owners. We extend such schema by enabling data providers to submit multiple cryptograms in a time window, by using time-varying encryption keys. We evaluate the impact of the noise component to the generated cryptograms so that to evaluate the data loss during the encryption process.