Research On Non-temporary Query Location Privacy Protection Algorithm Based On Prediction

Owing to the popularity of mobile network and the soaring trends of cloud computing, location-based service(LBS), in which users can utilize the geographical information for gaining entertainment services, has aroused widely attentions. In LBS, users with location-aware mobile devices are able to make queries about their surroundings anywhere and at any time. However, users have to upload their current exact location information to the location servers to get related services. This action will disclose user’s current location which may harm the privacy of the user. A lot of researches have been conducted concerning how to enjoy location-based services while protecting the location privacy of mobile users. To solve this problem, many different methods have been proposed, the most kind of which is K-anonymity method. The K-anonymity method form anonymous area containing K users to replace the exact location. However, existing techniques compute a node’s cloaking region based on the locations of its neighbors and the existing algorithm cannot judge whether the current result is the best.Given that, from a new perspective, we design novel query privacy preserving algorithm named prediction-based non-temporary query(PBNQ) in this paper. In light of this observation, we propose a concept of users’ potential location. The potential location is defined to be a user’s possible location. This algorithm is based on reasonable prediction, which considers the client’s moving status and trend- the velocity and direction- when generating cloaked regions. On the basis of predicting the user’s location, we propose an algorithm to predict and get the best time to form the anonymous region for better service. At first, the algorithm predicts users’ possible location. It calculates the anonymous area based on K-anonymity and determines whether the current moment is the optimal time.When the query users arise to very large, the server must be under great calculate pressure. At the same time, the relationship of spatial-temporal among the users has not been fully used for anonymous. To tackle the above issues, we propose an improved multi-user cooperation model and algorithm(MC-PBNQ).The architecture consists of trusted central servers and distributed servers. We propose a concept of parallel timeline. The algorithm will add or subtract users in the predict progress to get more accurate location information, which to predict the area better and improve the accuracy of location service.By analyzing the experimental simulation, we find that the probability based predictable query algorithm can predict the optimal time and guarantee a high accuracy. It improves the quality of services. On the basis of PBNQ, the cooperation algorithm improves the computational efficiency and the accuracy of location service.