ESPPTD:An efficient slicing-based privacy-preserving truth discovery in mobile crowd sensing

Abstract

The truth discovery approach can effectively obtain ground truth from the conflict sensing data provided by multiple participants, but privacy leakages restrict the enthusiasm of users for participating in the MCS, ESPPTD:an efficient slicing-based privacy-preserving truth discovery is proposed, which updates user weights and evaluation object truth values through secure iterations. ESPPTD first divides users into several clusters according to their location and number. The nodes in each cluster divide the sensing observation data into two parts according to the Extended Euclidean algorithm, and one part of the data is sent to any other node in the cluster. Secondly, the node mixes the received slice data with the remaining slice data and sends them to the cluster head node to ensure the privacy. Finally, the cluster head merges the slice data of all nodes in the cluster and uploads it to the server, and the node completes the weight update according to the aggregation result broadcast by the server, and the server cannot obtain the sensing observation data and weight of a single user. The truth evaluation is performed in the same way. In addition, in order to further reduce the user’s communication overhead and enhance the practicability of the system, this paper proposes an improved slicing-based privacy-preserving truth discovery in mobile crowd sensing based on ESPPTD, which eliminates slices. The forwarding requirement reduces the amount of computation and communication while ensuring data privacy. Experimental results show that ESPPTD can identify true and reliable data information while protecting data privacy.