IFTS: A Location Privacy Protection Method Based on Initial and Final Trajectory Segments

Abstract

Privacy protection problem is one of the most concerning issues related to Location-Based Services (LBS) in our daily life. Privacy protection of LBS often requires anonymizing customer's trajectory data. Currently available methods for trajectory anonymity often assume an entire trajectory as one anonymous unit, which may lead to low anonymity efficiency due to the massive amount of trajectory data, especially for customers travel through a long road. Considering people's routine activities, the starting and ending locations of a trip often uncover the user's request intent, which may lead to exposure of user privacy. In order to address the problem of inefficient trajectory anonymity, we propose a location privacy protection method that is based on the initial and final trajectory segmentation (IFTS) in this paper. In the IFTS method, the road network structure is first transformed into an edge cluster model based on its location type. Then, the user trajectory is divided to segments according to the temporal sequence of the ingress and egress nodes. The initial and final trajectory segments are identified and divided into equivalence classes, which then are used for constructing a trajectory graph and the corresponding k-anonymous set. Our experimental results show that the proposed method can reduce the anonymous area, improve anonymity efficiency, and enhance trajectory data utilization compared to the existing methods for trajectory anonymity.INDEX TERMS Location based service (LBS), location privacy, k-anonymity, initial and final trajectory segments (IFTS).

cFL：Data distribution edge clustering algorithm based on deep Federated learning

“…Mobile edge nodes installed in each LAN gather and aggregate essential or non-malicious local updates during the edge aggregation phase because each global aggregation uses network connection and cloud computing resources. Once the edge aggregated models have been delivered to the cloud server for global aggregation, delay associated with high concurrent access to the server is avoided, and the utilization of the computing resources of the mobile edge nodes is rationalized to relieve the load on the cloud [12]. The main contributions of this paper are as follows:…”