移动社交网络中面向多维隐私保护的安全查询方法研究

Over the past decade, mobile social network (MSN) services have been rapidly developing and are changing people's life style constantly. MSN is an open-platform with complicated user relationships, which then will cause a crisis of confidence and privacy disclosure. This problem has become the bottleneck for sound development of MSN. Preserving user privacy and avoiding information disclosure, therefore, need serious concerns, in the process of promoting the performance of service. This proposal intends to study the secure query method with multi-dimensional privacy protection in MSN, considering the complexity of the mobile network environment. We focus on popular applications and services in MSN, such as interest of point (POI) query, proximity detection, and friend matching. Firstly, we will analyze the problem of correlation attack existing in POI query, and propose a Privacy Preserving and Content Protection (PPCP) method for the continuous query based on the location transformation and public key cryptography; secondly, to satisfy the users' individual requirements, we will design a Flexible and Private Proximity Detection (FPPD) model to efficiently realize nearby friend searching in different scale of vicinity regions; finally, to preserve user’s location privacy and attribute privacy, we will present a Privacy-aware Fine-grained Friend Matching (PFFM) method to perform fine-grained management on different user attribute and to achieve multi-dimensional privacy protection for mobile users. The expected outcomes of this proposal will be helpful for realizing the user privacy protection in mobile social network.

近年来移动社交网络（MSN）服务迅猛发展，正不断改变着人们的生活方式；但由于MSN存在平台开放性、用户关系复杂性等特点，信任危机与隐私泄露问题已成为MSN健康发展的瓶颈。如何在提升服务水平及质量的同时全面保护用户隐私成为了亟待解决的问题。为此，本项目综合考虑移动网络环境的复杂性，针对MSN中的典型应用和服务，开展移动社交网络中面向多维隐私保护的安全查询方法研究。首先分析兴趣点查询中的关联攻击问题，利用位置空间转换及公钥加密技术，研究连续查询请求中面向用户位置和内容的隐私保护方法；其次根据用户个性化需求，基于网格哈希和保序加密技术，设计可扩展的私密好友近邻查询方案，实现两级邻域范围尺度下高效的近邻检测；最后综合考虑用户的位置隐私和属性隐私，结合细粒度的用户特征属性权重分级思想，研究细粒度相似好友查询的多维隐私保护方法。项目预期研究成果可为未来移动社交网络应用中的用户隐私保护提供借鉴。

近年来移动社交网络（MSN）服务迅猛发展，正不断改变着人们的生活方式；但由于MSN存在平台开放性、用户关系复杂性等特点，信任危机与隐私泄露问题已成为MSN健康发展的瓶颈。如何在提升服务水平及质量的同时全面保护用户隐私成为了亟待解决的问题。鉴于此，项目针对MSN中的典型应用和服务，开展了移动社交网络中面向多维隐私保护的安全查询方法研究。首先，分析了兴趣点查询中的关联攻击问题，利用位置空间转换及公钥加密技术，提出了连续查询请求中面向用户位置和内容的隐私保护方法；其次，根据用户个性化需求，基于网格哈希和保序加密技术，设计了可扩展的私密好友近邻查询方案，实现了两级邻域范围尺度下高效的近邻检测；最后，综合考虑用户的位置隐私和属性隐私，结合细粒度的用户特征属性权重分级思想，提出了细粒度相似好友查询的多维隐私保护方法。上述研究成果可为未来移动社交网络应用中的用户隐私保护提供借鉴。项目组在重要国际期刊和会议上共发表高水平学术论文11 篇，其中 SCI 期刊论文 8 篇（CCF B类文章4篇，CCF C类文章4篇，ESI 高被引论文2篇），EI论文3 篇。申请国家发明专利6项，获得授权专利3项。