Research And Application Of Accountable Privacy Preserving Scheme For Time-Space Distribution Mobile Crowdsensing Tasks

In recent years,with the development of wireless sensor networks and mobile Internet technologies,mobile crowdsensing has received extensive attention from academia and indus-try,and has been widely used in environmental monitoring,intelligent transportation,and other scenarios.These scenarios typically require users to collect specific sensing data at a given time and place,ie,and such sensing tasks have spatio-temporal distribution characteristics.The data collected by mobile crowdsensing often involves the privacy of users,and the disclosure of private data poses a serious threat to users.In addition,after sensitive information is protected,malicious users may abuse their privacy advantages and upload the sensing data multiple times for their own purposes,thereby affecting the sensing task results.Therefore,protecting user privacy and preventing malicious users become a contradictory goalLocalized differential privacy is based on the assumption that data collectors are not trusted,and can guarantee the privacy of data under any background knowledge attack,while the availability of data depends on the amount of data.Therefore,in the case of limited partic-ipation users,it is very practical to explore a localized differential privacy algorithm suitable for mobile crowdsensing scenarios and ensure high availability of data aggregation resultsIn view of the above problems,this paper has done the following work for the mobile crowdsensing task of space-time distribution1.This paper constructs an accountable privacy-preserving scheme.The scheme utilizes the nature of BBS+signature to implement user anonymous identity authentication,and imple-ments the accountability of malicious users based on public identity2.Considering the different roles of location and collected sensing attributes in the sens-ing task data analysis phase and the small number of users participating in the mobile crowdsensing task,the localized differential privacy policy based on random response is used to process the sensing data.Specifically,the location set is randomly selected,the noise attribute value is generated for the noise location,and the attribute value bit string is randomly flipped to implement data privacy protection of the user3.This paper proposes an efficient privacy protection data aggregation algorithm,which aggregates the private sensing data through three steps of position frequency estimation,noise removal,and regression fitting to obtain the results of sensing tasks4.In order to verify the effectiveness of the responsible privacy-preserving scheme,this paper implements a prototype system and simulates the mobile crowdsensing task with spatio-temporal distribution characteristics based on real datasets.A series of comparisons experi-ments and analysis are made under different differential privacy protection granularities.The experimental results show that the computational and communication overhead of our scheme is more advantageous in the mobile crowdsensing scenario,and the high-accuracy privacy data aggregation result can be obtained with a small number of user participation...