Research On Privacy Preserving Technology Of Internet Of Vehicles For Spatial Crowdsourcing Task Allocation

With the development of the Internet of Vehicles and the improvement of mobile devices,spatial crowdsourcing as a new service has received widespread attention from academia and industry.Spatial crowdsourcing takes advantage of the vehicles’ resources and plays an outstanding role in ride-hailing,traffic road monitoring,map mapping,and other intelligent transportation.Spatial crowdsourcing task includes a target location that requires vehicles to travel there and completes the task,vehicles submit accurate location information to a third-party server for task allocation.However,the vehicle’s location information may disclose sensitive information such as the vehicle user’s home address,interests,and living habits.Therefore,how to protect vehicles’ privacy in the task allocation of spatial crowdsourcing has become an urgent problem.Cryptographic algorithms-based privacy protection technology needs to generate and distribute keys for vehicles online,which is not practical in reality.Resulting in a large number of tasks exceeding the set time.Furthermore,the privacy protection technology typically provides a uniform level of privacy protection for vehicles,resulting in insufficient or excessive privacy protection and affecting the utility of vehicle-submitted location data.Based on the shortcomings of the above work,the main research contents of this scheme are as follows:(1)In order to solve the problem that the task allocation scheme with privacy protection based on homomorphic encryption need to distribute key online,this scheme proposes a task allocation scheme based on distributed two-trapdoor public-key cryptosystem.This scheme completes task allocation under the ciphertext of task location and vehicles’ location that encrypted with different keys,which ensure that the data of the task and vehicle locations is secure.In addition,the system does not need to distribute keys for vehicles online during task allocation,which reduces communication costs and improves the efficiency of task allocation.(2)In order to solve the problem of the privacy requirements of vehicles participating in spatial crowdsourcing during driving,this scheme proposes a privacy quantitative measurement model based on reinforcement learning that can accurately describe the relationship between vehicle trajectory,vehicle spatial crowdsourcing execution information,and privacy requirements.In addition,reinforcement learning is used to design an adaptive adjustment mechanism so that the privacy quantitative measurement model can dynamically adjust while vehicles participate in spatial crowdsourcing at different locations.(3)In order to solve the problem that the privacy protection technology provides the same level of privacy protection for vehicles,which causes low data utility,this scheme designs a spatial crowdsourcing task allocation scheme with personalized privacy protection in the Internet of Vehicles based on localized differential privacy,which uses the dynamic privacy quantitative measurement model to calculate the vehicle’s current privacy protection requirements.The relationship between privacy protection level and privacy budget value in localized differential privacy is mapped through the activation function,thus realizing personalized privacy protection.Furthermore,in personalized privacy protection,the scheme combines the optimal Laplace mechanism to achieve a balance between data utility and privacy protection.