Research On Privacy-preserving Task Assignment In Spatial Crowdsourcing

With the popularization of 4G networks and the deployment of GPS,smartphones are equipped with more powerful hardware and people with smartphones can perform some geolocation-based tasks such as photo collection,traffic condition monitoring,and timely reporting in hot spots.Thus,a new crowdsourcing model called spatial crowdsourcing began to appear.The user who has smartphones is the worker of spacial crowdsourcing.In spatial crowdsourcing environment,it assigns workers to a designated location,and workers should physically move to this place to complete the task.Spatial crowdsourcing can effectively collect and process spatial data,which greatly benefits people's daily life.Therefore,spatial crowdsourcing has received a great deal of attention,and task assignment in spatial crowdsourcing has become a research hotspot.A large number of previous studies only pay attention to the quality of task assignment.While there is little research on the privacy protection of spatial crowdsourcing.In the spatial crowdsourcing execution procedure,users need to upload their privacy data(such as personal location,trajectory,etc.)to spatial crowdsourcing server(SC-server).Through analyzing users data operation,the malicious adversary can infer the personal information of the relevant users,such as home address,personal habits,and other personal privacy.It is a great threat to spatial crowdsourcing user's personal and property safety.Therefore,research on the privacy protection of spatial crowdsourcing task assignment is imminent and in this article,we mainly discuss the following two kinds of common privacy crowd protection scenes:In privacy-preserving velocity-aware spatial crowdsourcing task assignment,we consider the two points: 1)how to protect the privacy of both workers and the task is not invaded by attackers.2)how to improve the task assignment accuracy and reduce computation overhead under the premise of privacy protection.To realize the two goals,we propose a privacy-preserving protocol.With this privacy-preserving protocol,we can effectively protect both worker and task privacy.The protocol not only satisfies the requirements in terms of privacy and security,it also can guarantee the accuracy and the efficiency.We prove that the proposed protocol is privacy-preserving against semi-honest adversaries.Simulation on two real-world datasets shows that the proposed protocol is more effective than existing solutions and can achieve mutual privacy-preserving with acceptable computation and communication cost.In privacy-preserving downwind spatial crowdsourcing task assignment,we mainly consider the two points: 1)how to reasonably assign tasks to suitable workers to make sure the high completing probability and low cost in the downwind-type spatial crowdsourcing.2)how can we ensure the security of private data of the workers participated in this spatial crowdsourcing scenario.In this paper,we propose a privacy-preserving framework to protect workers privacy data during the task assignment procedure.We first prove that our spatial tasks assignment problem is NP-hard and give a corresponding greedy algorithm to solve it.Next,we give the performance analysis and security analysis of our privacypreserving framework.Both theoretical analysis and experimental verification prove that our proposed technique is quite effective and settles the problem nicely.