Research On Location Privacy Protection For Crowd Sensing Network In Smart City

The smart city has become a new direction of the global urban development.The popularization of the Internet and the development of mobile crowd sensing technology provide more guarantee for the realization of the smart city.In the smart city-oriented crowd sensing networks,the city users perceive the city data with mobile terminals and upload the perception results to the server,thus to provide the data source for those city managers to design scientific urban development plans.For example,the traffic management departments can release the real-time road condition information to help drivers to choose the optimal route planning.Also,by analyzing the track information,the urban planning department can plan the road and optimize the urban traffic network.Moreover,the developers can make business decisions through the crowd density analysis and behavior analysis,and then develop different functional businesses in different locations.It should be mentioned that,during the users' sensing process,users are often required to upload the location information in the city,which may result in the disclosure of users' privacy information,including users' addresses,the living habits,the interests and so on.How to protect the location information privacy for the sensing users has become an urgent problem in the crowd sensing networks for smart cities.This paper analyzes the location privacy disclosure problem in the crowd sensing networks for smart cities,and considers two different scenarios: the single location scenario and the trajectory location scenario,and designs the corresponding location privacy protection algorithms.The main purpose of this paper is to improve the availability of the released location data on the premise of ensuring their location data privacy.The main work of this paper includes the three parts shown as follows.(1)Aimed at the location privacy disclosure problem in the single location application scenario,this paper proposes a location privacy protection algorithm based on the differential privacy k-means(L-DPk-means).With the introduction of the concept of outliers,the users' location data sets are classified according to the crowd density function.The users' location data set is divided into two parts: the outlier location set and the non-outlier location set.With the use of the geographical inseparability method,the data in the outlier set are noise-added separately.For those data in the non-outlier location set,taking into account the characteristics of the users' location data in crowd sensing networks,and one improved k-means clustering algorithm for selecting the initial center point is proposed,then the number of clustering iterations is reduced,and the errors caused by the added noises are decreased too.(2)Considering that in the continuous location application,the L-DPk-means algorithm will consume a large amount of privacy budget,and also will result in poor availability of the trajectory data set,one improved DPk-means-based trajectory privacy protection(T-DPk-means)algorithm is proposed in this paper.The selection of the initial clustering center points of differential privacy k-means algorithm is improved,with the introduction of the trajectory density.According to the characteristics of the vehicle trajectory data and those of the road networks in smart cities,the concept of trajectory turning point is introduced.With the design of the turning angle measurement method,together with the design of the threshold in determining the turning points,the turning points in the trajectory data set are found.Then,the trajectory information of vehicles can be represented by continuous turning points.By weighting the turning points with high-density parameters,those turning points are more likely to be chosen as the initial clustering centers,to improve the availability of the published trajectory information.(3)Based on the aforementioned two location privacy protection scenarios,some experiments are designed with the simulated crowd sensing network environment in smart cities,and the effectiveness of the privacy protection algorithms L-DPk-means and T-DPk-means proposed in this paper is verified.Aimed at the L-DPk-means location privacy protection algorithm,one location-based services(LBS)application scenario is designed,where users can obtain the surrounding services by sending their position information.The experiment results show that the user's location deviation after privacy protection is small and has little impact on the service quality.When it comes to the trajectory privacy protection algorithm T-DPk-means,the simulation experiments are designed within the internet of vehicle application scenarios.The experimental results show that,with the use of T-DPk-means method,the availability of the trajectory data gets improved.