Research And Application Of Verifiable And Privacy-preserving Data Aggregation Scheme In Mobile Crowdsensing

In recent years,smart mobile terminals have rapidly increased in terms of processors,embedded sensors,and network transmission rates.By combining the idea of data crowdsourcing and the sensory capabilities of smart mobile terminals,a new data collection paradigm named mobile crowdsensing is proposed,which is fast,simple and low-cost.The mobile crowdsensing platform provides data support for various urban monitoring applications by recruiting a large number of mobile users to collect sensory data,which is widely used in the scenarios such as environmental monitoring,traffic monitoring,and medical health monitoring.However,the cloud platform is untrusted and there is a risk that the cloud platform may leak the privacy of mobile users during the data aggregation process.In addition,the cloud platform may provide false calculation results to trick the task publisher in order to save its own computing resources.In mobile crowdsensing,providing task publishers with a mechanism to verify the correctness of computation results and building an appropriate privacy-preserving mechanism for mobile users are the focuses of this thesis.In an untrusted cloud platform computing environment,the following work in mobile crowdsensing for privacy leakage and verifiable computation results is investigated:1.This thesis constructs a verifiable privacy-preserving data aggregation scheme.The scheme uses data masking technology combined with the nature of bilinear mapping to achieve privacy-preserving multi-dimensional data aggregation and verifiable computation.Meanwhile,the problems of missing mask value caused by mobile user exiting are solved by using Shamir's secret sharing algorithm.2.This thesis considers the task situation of continuously collecting multi-dimensional sensory data in the periodic internal location area,and constructs the multi-dimensional sensory data into a composite data through the Chinese remainder theorem.Consequently,the privacy of the multi-dimensional data aggregation in the location area,the mobile user location and sensory data privacy are protected so that two types of collusion attacks can be resisted.3.This thesis proposes an efficient verifiable calculation and privacy-preserving data ag-gregation algorithm.The task publisher verifies the correctness of the cloud platform's calculation results through verifiable public keys and calculated credentials.4.In order to show the effectiveness of the verifiable privacy-preserving data aggregation scheme,this thesis implements a prototype system,and two similar schemes on the real data set(the privacy-preserving multi-dimensional data aggregation scheme and verifiable privacypreserving data aggregation scheme)are compared in the simulation experiments.The experimental results show that our proposed scheme is more suitable for the scenarios where data reports are submitted continuously during the interval period.Also,the calculation overhead and communication overhead are more advantageous,and it is efficient in the computational cost of verifying the correctness of the aggregation results.