在群智感知网中提供差分隐私保护的研究  

A differential privacy protection is provided in the group intelligence perception network

作　　者：刘倩[1] 张可佳 李可扬 LIU Qian;ZHANG Ke-jia;LI Ke-yang(Harbin Institute of Petroleum,Harbin 150001,China;School of Computer Science and Technology,Harbin Engineering University,Harbin 150001,China;School of Computer Science,Wnhan University,Wnhan 430070,China)

机构地区：[1]哈尔滨石油学院,哈尔滨150001 [2]哈尔滨工程大学计算机科学与技术学院,哈尔滨150001 [3]武汉大学计算机学院,武汉430070

出　　处：《信息技术》2018年第9期134-138,共5页Information Technology

摘　　要：群智感知网作为最先落地的物联网在社会上引起了广泛的关注。由于用户在群智感知网中要共享自己的位置信息,个人隐私的泄露是群智感知网在普及中面临的巨大问题。在群智感知网中经常面临的一种查询就是k NN(k Nearest Neighbors)查询,即给定一个任务位置点,查询距离该位置点最近的k个用户。文中为群智感知网中的k NN查询设计了一种隐私保护机制。通过对查询结果做简单合理的随机扰动,能够满足ε-差分隐私,从而为群智感知网中的k NN查询提供了最强有力的隐私保护,消除人们对群智感知网中隐私泄露的顾虑。实验结果表明该隐私保护机制具有较好的数据可用性。As the first Internet of things to be launched,group intelligence perception network has attracted wide attention in the society. Since users should share their location information in the group intelligence perception network,the disclosure of personal privacy is a huge problem faced by the group intelligence perception network in its popularization. A query that is often faced in the group intelligence perception network is the k NN（ k Neighbors Nearest） query,that is,given a task location point,the Nearest k users are queried. This paper designs a privacy protection mechanism for k NN query in group intelligence perception network. Through the query results to do a simple and reasonable random disturbance,privacy,can satisfy the epsilon-difference for k NN query in the group of mental perception network provides the most powerful privacy protection,eliminate people’s group of mental perception net privacy concerns. The experimental results show that the privacy protection mechanism has good data availability.

关 键 词：群智感知网 差分隐私保护 KNN查询 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]