作　　者： ; ; ;

机构地区： 东北大学软件学院

出　　处： 《通信学报》 2015年第S1期94-102,共9页

摘　　要： 随着移动设备和定位技术的发展,产生了大量的移动对象轨迹数据,相伴而来的是个人隐私泄露问题。现有的轨迹隐私保护研究均假设轨迹数据是准确无误的,但由于数据采集设备不精确、移动对象延迟更新等原因,轨迹数据不确定性普遍存在。提出了一种基于K-匿名的不确定轨迹数据隐私保护方法,对发布的数据进行隐私处理,该方法首次将线性轨迹转化为不确定区域的思想引进轨迹数据的隐私处理。首先,使用概率统计的方法将轨迹泛化成一个更为真实的轨迹区域,然后将相似度高的轨迹域聚合成等价类进行数据的隐匿和发布,最后在真实的数据集上进行实验。 With the development of location based service(LBS) and location-aware devices,the amount of trajectories of moving objects collected by service providers was continuously increasing, meanwhile, it can cause great threaten for personal privacy. Most researches of trajectory privacy preserving were on deterministic data, however, trajectory's uncertainty was inherent due to the inaccuracy of data acquisition equipment, delayed update, and so on. A new method was prosed to protect the privacy of trajectory data in publishing. It is the first time to present the idea that transforming the trajectory to an uncertain area to cluster. First, a probability statistics method to model the trajectory to an uncertain area was proposed. Second, the similar uncertain area into a cluster was put and sanitized in an equivalence class. Finally, the performance of the proposal was compared with(K δ)-anonymity model in real datasets.

关 键 词： 轨迹数据发布 隐私保护 不确定性轨迹 匿名 轨迹聚类

领　　域： [自动化与计算机技术] [自动化与计算机技术]