Location Privacy Protection Approaches Research In Mobile Internet

In recent years, the growth of the Mobile Internet is very rapid. The Location Based Services is used widely in academia and industry brings great importance of national science and technology departments. Geo-spatial information industry, such as the Location Based Services, has been an important part of Mobile Internet. It is closes contact with the national economy and social development, improving the people’s living standards and making daily life more convenient.However, people enjoy the unprecedented convenient brought by there location services, the challenges of privacy information security become more and more urgent. Today, users already believe that the privacy risk of sharing and publishing their location information outweigh the profit in many location services. How to effective control the useage of location information, while at the same time looking after their QoS of location services and privacy security, have become key topics in privacy protection and location services. From the privacy security of the mobile terminal, focusing on the topic of location access and publish control for the applications, this dissertation addresses several important problems such as the location and trajectory anonymous in location query services, trajectory data analysis and privacy metric, application’s behavior analysis with its emploied in Android. Besides, we present several solutions and real-word systems based on the real trajectory data and application background in this dissertation. Specifically, the main results of this dissertation are summarized as follows:1. We present a location anonymous algorithm based on Multi-POI Feedback. Existing approaches on location k-anonymous are mainly based on POI feedback by one time’s, which cannot apply on the online query services because of the calculation costly. Based on the user-defined privacy requirements and quality of service and trusted third-party server, our method employs the recursive quadtree partition retrieval to stepwise selection objective point of interest, hence, enable an efficient privacy protection.2. We present a background knowledge attack oriented multi-dimensional trajectory anonymity algorithm. Focusing on the background knowledge attack based on the spatial-temporal correlation, this method is built on the anonymous trajectory data collection system and achieves the protection of user privacy through spatial-temporal anonymity algorithm in the multi-user collaboration privacy protection mode. Based on the real trajectory data, the experimental results demonstrated that this method can prevent the background knowledge attack based on the position and moving mode effectively and meets the demands of the privacy protect of k-Anonymity.3. We present a spatio-temporal trajectory anonymous algorithm via map matching. Focusing on the current approach using spatio-temporal trajectory anonymous neglected the influence of the trajectory on the spatial mapping, and a spatio-temporal trajectory anonymous algorithm via map matching is presented to meet the demand for security of the spatio-temporal anonymity and the spatial correlation in this paper. The scheme first makes a candidate matching points set for dummy trajectory points by use of the real road network datasets, and then searches the target match trajectory points based on the anonymous function of temporal and spatial and the heuristic breadth first search. Experiments validate the efficiency of the privacy protection algorithm by use of a real datasets and a tool of spatio-temporal trajectory anonymous via map matching that is designed and implemented based on Android platform.4. We present an action trajectory prediction and publish control algorithm. Focusing on the current privacy protection methods ignore that the temporal and spatial correlation of location points has important impact on the leakage of sensitive information. In this work, our primary contributions are (1) to advertise a risk model based on restoring sensitive points in the trajectory, and (2) to construct a prediction model for trajectorys based on Markov chain, and (3) to calculate the conditional probability of restoring sensitive points in the trajectory based on the context information of time and space. We ran an extensive set of experiments on large real-world trajectorys. Our results demonstrate the possibility of restoring location points with our algorithm is above80%. Further, to advertise the issue of privacy leakage, we also propose a metric model. By calculating the difference value of the prior probability and the posterior probability of sensitive point, it measures the degree of privacy leakage resulting from restoring location points. We use the metric model to evaluate the influence of the issue of the trajectory points on sensitive points. Our experiments demonstrate the effectiveness of our privacy protection method.5. We present and implement a location services access control system for Android. Focusing on the complexity of location services application and unity of privacy protection method, we present an adaptive location privacy protection method based on context-aware, in view of the type of application refer to currently the environment and their own behavior. Simultaneously, we present a method to intercept and modify user’s location for Android. Based on the real trajectory data, we present sensitive position detection and clustering algorithm. Identify the user from the track position in the time domain, and using probabilistic model to analyze the risk of the user divulging sensitive position. In this paper, the experimental whit the real data, the results verify the validity and reliability of the system.