Research On Secure Mobility Prediction And Incentive Mechanism In Spatial Crowdsourcing

The continuous development and popularization of spatial crowdsourcing not only bring convenience to people’s lives,but also arouse people’s concern about privacy and service.On the one hand,location information is likely to be over-collected when users use applications.The mobility data composed of users‘locations are not only mined and utilized,but also have the risk of privacy disclosure.On the other hand,malicious task participants may collude with each other in an attempt to cheat other participants to obtain more benefits.As a result,fewer and fewer workers are willing to complete tasks,which has an impact on the service quality and further affects the whole crowdsourcing ecology.Therefore,it is worth considering how to provide effective service while protecting the privacy of users,realize the identification and resistance to malicious participants,and motivate participants’ enthusiasm.This paper analyzes the privacy issues,malicious collusion,and low enthusiasm of participants in spatial crowdsourcing and its applications.The main work and contribu-tions of this paper are as follows:1.Aiming at the problem of user mobility privacy in spatial crowdsourcing applica-tions,and considering the risk of privacy disclosure of neural network parameters trained by service providers,we propose a privacy- preserving mobility prediction scheme for spatial crowdsourcing.Firstly,based on secret sharing technology,we design several privacy- preserving computing modules between two parties.On this basis,we propose a set of privacy protection interaction protocols for different stages of the prediction network.Protocols allow the client and server to complete the mobility prediction in a cooperative way on the premise that their private data is not leaked.Theoretical analysis and experiments show that the scheme not only protects user mobility data and service provider’s prediction model parameters,but also ensures the accuracy of prediction results.2.In view of the potential malicious collusion and low enthusiasm of participants in the process of spatial crowdsourcing,we propose an incentive mechanism to re-sist malicious participants in spatial crowdsourcing.Firstly,we design a reputation calculation model based on the traditional multi-weight subjective logic model.On the one hand,it improves the influence of malicious interaction on reputation and realizes the identification of malicious workers.On the other hand,the trust value between task publishers is considered,which is used as the basis for calculating the recommendation credibility to realize the identification of malicious task publish-ers.Then,based on the adverse selection model and moral hazard model in contract theory,different incentive mechanisms are designed according to the characteris-tics of new and veteran workers.Finally,the simulation verifies that the designed incentive mechanism can resist malicious collusion,which ensures the overall re-liability of participants and stimulates participants’ enthusiasm by improving the benefits of both parties.