Research On Differential Privacy Protection Methods In Spatial Crowdsourcing Environmen

In the crowdsourcing mode,people distribute tasks to non-specific workers through the crowdsourcing platform,and space crowdsourcing requires workers to arrive at the task site to complete the task,so it is expected to distribute the task to workers who are close to each other.Traditional spatial crowdsourcing is that users report their geographical location to the platform,and the platform performs task matching according to factors such as distance.However,the direct exposure of user geographic location information to the spatial crowdsourcing platform may lead to privacy disclosure.In order to protect the privacy and security of users in the space crowdsourcing environment,two location privacy protection schemes for space crowdsourcing users are proposed.The two schemes protect the location privacy of space crowdsourcing users under different spatial crowdsourcing system models,while ensuring the effectiveness of task allocation.The main work of this paper includes the following:(1)This paper proposes a spatial crowdsourcing location privacy protection scheme which satisfies the centralization of differential privacy.In the traditional spatial crowdsourcing system model,a trusted central server is introduced to collect workers’ location information and focus on privacy protection.In this paper,a worker privacy space decomposition(PSD)algorithm is proposed.firstly,the worker location is adaptively meshed;then,the data dispersion is measured based on the radius of the standard deviation circle,the privacy budget allocation is optimized,and the data utility is improved;finally,the theory proves that the algorithm satisfies differential privacy.(2)Aiming at the problem that the PSD of differential privacy protectors will reduce the efficiency of task allocation,this paper proposes a new grid-based greedy algorithm to construct the optimal task broadcast area(GR),which ensures the task acceptance rate of GR and improves the effectiveness of task allocation.The experimental results show that the utility of data and the performance of task allocation are improved under the same privacy budget.(3)This paper proposes a spatial crowdsourcing location privacy protection scheme that satisfies local differential privacy.Under the system model that does not rely on a trusted central server,it not only protects the user’s location privacy but also protects the effectiveness of task assignment.First,a Hierarchical Separation Tree(HST)structure is introduced to construct a set of location points into an HST;then,a local differential privacy mechanism based on HST is designed to perturb the location nodes at the local end,and the mechanism is proved theoretically Geographic indistinguishability is satisfied;finally,experiments verify that under the same privacy budget,this scheme outperforms existing differential privacy mechanisms in terms of the total distance of task assignment.