一种实时轨迹隐私保护策略

摘要 实时轨迹隐私问题是LBS（Location-BasedServices）领域的一个重要问题.虚假轨迹技术是一种流行的隐私保护技术，它产生多条与真实轨迹相似的虚假轨迹.然而，已有的虚假轨迹保护技术并未考虑用户所处的实际环境以及相邻时刻的位置关系等约束，从而使得攻击者很容易借助其他背景知识推测出用户的真实轨迹.因此，本文在所提出的两种全新隐私保护算法中应用了信息熵和位置可达性约束，这两种算法分别为虚假轨迹生成DTG（Dummy-Based TrajectoryGenerating）算法、增强型虚假轨迹生成EnDTG（Enhanced-DTG）算法.实验结果表明，相比于现有方案，本文所提的方案能有效保护用户的轨迹隐私.
Abstract：Real-time trajectory privacy protection is a key research topic in the field of location-based services (LBS). Dummy trajectory technology is a popular privacy protection method which generates multiple fake trajectories similar to real ones. However, existing schemes take into account neither the real environment nor the relationship between adjacent positions carefully; with the help of this ancillary information, the real trajectory can be inferred easily. Hence we integrate entropy and constraints on location reachability into our proposed schemes, named dummy-based trajectory generating (DTG) and enhanced-DTG (EnDTG). Experimental results show that both schemes offer a higher privacy level than existing ones.