| Mobile Crowd Sensing(MCS)is an efficient and low-cost data collection model that has received a lot of attention in recent years.It uses a large number of user smart devices to sense the physical world and obtain various information such as personal behavior,device status,and surrounding environment.User selection and incentive are two essential parts of MCS,which can select users that meet the needs of the sensing task and provide a continuous and high-quality data source for the sensing platform.However,the proper execution of both processes requires users to upload a large amount of data containing private information,putting user privacy at risk of leakage.The thesis addresses the privacy leakage problem of MCS user selection and incentive scenarios,the main work of the thesis is as follows:(1)A privacy-preserving user selection mechanism for spatiotemporal preferences is proposed for the MCS user selection scenario.It adopts the user-task matching approach to achieve user selection and defines the Time Matching Score(TMS)and Location Matching Score(LMS)of users and tasks for the spatiotemporal requirements of MCS.In addition,considering that there is still room for improvement in matching efficiency and granularity in existing research,we construct a lightweight Secure Calculation Protocol(SCP)based on Shamir secret sharing and Carmichael Theorem to achieve secure calculation of TMS and LMS as well as secure matching of attribute value size and range.Our detailed theoretical analysis and experimental tests verify that this mechanism can protect the privacy of both sensing users and task publishers and has high efficiency.(2)For the incentive scenario of MCS,we propose truth discovery and incentive mechanism based on secure multi-party computation.It can calculate the data ground truth for data requesters and dynamically allocate user rewards based on the weights calculated from the truth discovery.To address the privacy leakage in truth discovery,we construct a privacy-preserving mechanism based on additive secret sharing and the Sharemind secure multi-party computing platform,which achieves efficient and privacy-preserving truth discovery and eliminates the potential of data misuse.In addition,to motivate users to provide high-quality data,we design a dynamic reward calculation method using the user weights generated by the truth discovery for a data quality-driven incentive mechanism.We verify through detailed theoretical analysis and experimental tests that this mechanism can provide protection for data privacy and still have high execution efficiency in scenarios with a large number of users and tasks. |
