Research On User Differential Privacy Protection Mechanism In Spatiotemporal Crowdsourcing

With the continuous development of Internet network technology,various online applications based on the Web have developed rapidly in recent years,which also provides a new way for people to gather collective wisdom,crowdsourcing,to solve problems that computers cannot solve independently.Due to the rapid development of technologies such as the mobile Internet and the Internet of Things,the data management technology called crowdsourcing has transitioned from a business model based on online crowdsourcing to ‘spatio-temporal crowdsourcing’,also known as spatial crowdsourcing or mobile crowdsourcing.With the more complex and diverse needs,time-space crowdsourcing has gradually appeared in people’s field of vision.How to protect the security and privacy of spatiotemporal crowdsourcing while reducing the impact on service quality is an important research direction in the field of spatiotemporal crowdsourcing.In response to the problems raised above,this thesis makes an in-depth analysis of the current privacy security technologies in spatial crowdsourcing based on the complex characteristics of time and space in space-time crowdsourcing,and discusses its advantages and disadvantages.This paper proposes a spatiotemporal crowdsourcing user privacy protection model.In the k-anonymity technology,differential privacy technology is added to protect their real-time location data in the stage of crowdsourcing user perception data,which improves the privacy of spatiotemporal crowdsourcing users.The protection level is verified by simulation experiments,which effectively proves the effectiveness of the scheme for the privacy protection of users’ location in time-space crowdsourcing,better maintains the security of the network platform,and realizes the optimization of network services.On the other hand,due to the correlation of time,there may be a problem in only protecting the privacy of the user’s location data,that is,the attacker may infer the user’s daily trajectory through the user’s historical upload location,and then infer the user’s daily trajectory according to the time sequence.User’s uploaded private data.Therefore,it is necessary to increase the privacy protection considering the user’s upload time data to strengthen the user location privacy protection in spatio-temporal crowdsourcing.Therefore,this thesis further proposes a spatio-temporal data privacy protection scheme in spatio-temporal crowdsourcing.In this scheme,first,a differentially private spatio-temporal location generalization algorithm is proposed to eliminate the dependence of existing mechanisms on implicit assumptions.The algorithm performs temporal generalization on the original trajectories by merging similar times,and obtains a new time-similar spatio-temporal trajectory dataset.Next,the positions are probabilistically merged at the same point in time in the new trajectory dataset using an exponential mechanism.Furthermore,since this mechanism prefers to incorporate locations belonging to closer trajectories,it can prevent serious side effects on data utility during generalization.Then,a differentially private publishing algorithm for spatio-temporally similar trajectories is proposed to publish the merged trajectories.A distance constraint between two consecutive locations is also introduced to guarantee the utility of the data.At the same time,it is proved that the scheme satisfies differential privacy,and a privacy protection strength algorithm is proposed to verify the security of the scheme.