Research On Privacy Protection Strategy Of Attribute Relationship Evaluation In Crowd Sensing

Mobile Crowd Sensing(MCS)uses mobile terminal equipment carried by users as the basic sensing unit,and relies on a large number users to form a large-scale sensing network,thereby completing some complex sensing tasks that are difficult to be achieved by individuals.These sensing users who carry various smart sensors upload the collected sensing data to the sensing platform.The platform combines data analysis and mining technologies to extract favorable values.Although perceptual users rely on a large number of advantages to collect enough data to solve large-scale social problems in a fast,easy and cost-effective manner,users' personal information is vulnerable to threats in the process of task allocation and data reporting.Many users will be reluctant to participate in the data collection process without appropriate privacy protection mechanisms.Therefore,it is important to design an effective privacy protection mechanism.First,this thesis introduces the research background and the system architecture of crowd sensing networks,summarizes the characteristics and research hotspots of the crowd sensing networks,and then analyzes the characteristics of privacy protection in crowd sensing networks,and discusses in detail the challenges faced by privacy protection during task assignment and data reporting,and introduces typical crowd sensing privacy protection methods.Secondly,a collaborative mechanism of task assignment and privacy protection based on user attribute sensing is proposed to solve the problem of sensing task matching and protecting privacy information of sensing users in mobile crowd sensing network.First,the static attributes and social attributes of users are predicted by mining users' different tendencies,willingness and access to tasks based on the perception of users' inherent attributes and the participation records of historical tasks.Then,combined with the user attributes as input,BP neural network is used to analyze the user's service ability,so as to realize the optimal assignment of tasks and users.Finally,the sensing user generates pseudonyms to participate in the sensing task,and combines the ring signature to generate random numbers for user attributes to encrypt the attributes,ensuring that sensing users improve the accuracy of the platform's perceptual data under the premise of privacy and security.Simulation results show that the proposed mechanism can effectively select sensing users,verify the availability of uploaded data by user,and protect the identity security of users.Then,a task attribute privacy user matching scheme based on edge server is proposed,aiming to solve the problem that the sensing platform processing a large amount of user information increases the time delay and the high bandwidth cost and leaks user information.Firstly,the task requester sends the task requirement to the sensing platform,and then releases the task after sensing platform scalar task attribute.Then,the sensing user generates the user task attribute based on the task attribute and uploads it to the edge server.Secondly,the edge server uses the weight clustering method to group user task attributes,and the edge server is used as a group management to introduce a group signature to the user and upload it to the sensing platform.Finally,the sensing platform efficiently calculates the attribute similarity under the condition of the user group signature,selects the participating user set based on merit,returns the result to the edge server,and the edge server notifies the real user to upload the sensing data for signature verification.Simulation results show that the proposed mechanism can securely select sening users and reduce network delay.Finally,the work of this thesis is summarized and the future work is prospected.