Abstract: The attribute of the user can be collected by the adversary as background knowledge, as the user utilizing the continuous query in location-based service. Then with the attribute revealed along each query in the continuous request, the adversary can get the location privacy of the user with attribute correlation. In order to cope with this problem, a particle swarm optimization clustering scheme is proposed. This scheme can both accelerate the procedure of similar attributes finding and preserve the location privacy of the user. In this scheme, a central server is employed as usually assumed in the scheme of location privacy preserving. The central server performs the particle swarm optimization with the received anonymous request, and accelerates the process of choosing anonymous users with similar attributes. Then anonymous users with similar attributes will achieve the attribute anonymity and location generalization that obfuscate the correlation between each other, and then the attribute anonymity and location generalization make the adversary difficult to identify any special user with attribute correlation. At last, the experimental results verify the fact that, the particle swarm optimization scheme provides a better privacy preserving level and a faster processing speed than other similar algorithms.

Key words: location-based service, continuous query, particle swarm optimization clustering, attribute anonymity, location generalization, privacy preserving

摘要： 针对基于位置服务中连续查询情况下，用户自身属性信息很容易被攻击者获取，并通过关联获得用户位置隐私的情况，提出了一种利用粒子群聚类加速相似属性用户寻找，并由相似属性匿名实现用户位置泛化的隐私保护方法。该方法利用位置隐私保护中常用的可信中心服务器，通过对发送到中心服务器中的查询信息进行粒子群属性聚类，在聚类的过程中加速相似属性用户的寻找过程，由相似属性用户完成位置泛化，以此实现位置隐私保护。实验结果证明，这种基于粒子群属性聚类的隐私保护方法具有高于同类算法的隐私保护能力，以及更快的计算处理速度。

关键词: 基于位置服务, 连续查询, 粒子群属性聚类, 属性匿名, 位置泛化, 隐私保护