摘要
在这个移动互联网技术和大数据技术快速发展的时代,基于位置的服务使移动用户的位置信息数据化,给人们的生活带来极大便利的同时也威胁到了移动用户的位置隐私.传统的位置隐私保护方法只对当前位置和当前时刻的隐私进行考虑,这类方法既没有严格的隐私度量标准,又无法应对在攻击者得到用户的历史时序位置信息的情况下进行概率推测攻击.针对传统方法的这些问题,该文基于概率推测模型设计了一种位置隐私保护算法MaskK,首先通过隐马尔可夫模型(HMM)对用户的移动状态和位置发布情况进行建模,计算出用户移动位置的抑制发布概率向量,然后利用该概率向量中的概率对用户的部分位置进行抑制发布,使攻击者通过搜集到的用户历史位置数据得到的信息量尽可能的小,并引入k-匿名思想和粒子群优化算法(PSO)进行优化,进一步提高算法的运行效率和服务质量.该文通过真实数据对提出的算法进行了科学的实验,验证了MaskK在隐私保护效果、服务质量和运行效率上的优越性.
In the era of rapid development of mobile Internet technology and big data technology,mobile users can access a variety of location-based services through mobile smart devices,which becomes an indispensable part of people’s daily life.However,location-based services make the location information into data records,which brings a great convenience to people’s daily life,at the same time,also brings threats to location privacy of people.Further,this may be detrimental to the user’s personal and property safety.Traditional privacy protection methods only consider protecting location privacy of the current time and the current location,and this kind of methods don’t define strict mathematical model of privacy measurement.Moreover,if there is a probabilistic inference attack based on the user’s timing position data,these methods can’t work very well.To this issue,an improved location privacy protection algorithm MaskK is proposed,which is based on probabilistic inference.Firstly,we use Hidden Markov Model(HMM)and the history location data of the mobile user to establish a model,which shows the user’s mobile state and location releasing state.And we define a privacy metricδ-privacy,we can use it to measure the level of privacy by calculating the difference between the posterior probability and the prior probabilities of the user transfer according to the user’s location data released.Secondly,based onδ-privacy,in order to protect the user’s location privacy,we need to suppress the release of some of the user’s locations,so that the attackers get the posterior probability as little as possible by the user’s locations released.That is,according to the user’s new location data released,the attackers get the amount of information that is as little as possible.So,we need to compute the suppressed release probability vector of the user’s location.Meanwhile,we introduce k-anonymity and Particle Swarm Optimization(PSO).In the process of user locations release,the k-anonymity is used to guarantee the user’s sensitive locations which are difficult to be distinguished.So as to achieve a fuzzy for user’s sensitive location,which increases the difficulty of the attacker to guess the sensitive locations of the user.In the process of calculating the suppressed release probability vector,the PSO algorithm is used to improve the efficiency of the algorithm,which finds the probability vector that satisfies theδ-privacy requirement by multiple iterations.In addition,this paper defines a fitness function of the particle,we can find the relatively optimal suppression release probability vector to protect the user’s sensitive locations privacy by finding the extremum that satisfies the fitness function of the particle.Moreover,to prevent an attacker from intercepting the LBS request of user between the user and the trusted center server,we design a kind of user-centric system model.Thus the trusted server only needs to compute the release probability vector.Through the above process,we can make sure that the users’sensitive locations could be protected well,and the efficiency of algorithm is high.At last,we use the real data to conduct a scientific experiment for MaskK,which shows that the MaskK algorithm performs well on the effect of privacy,quality of service,and the efficiency of execution.
作者
李婕
白志宏
于瑞云
崔亚盟
王兴伟
LI Jie;BAI Zhi-Hong;YU Rui-Yun;CUI Ya-Meng;WANG Xing-Wei(Department of Computer Science and Engineering,Northeastern University,Shenyang 110169;Department of Software,Northeastern University,Shenyang 110169;Wireless Business Group,Qunar.Com,Beijing 100080)
出处
《计算机学报》
EI
CSCD
北大核心
2018年第5期1037-1051,共15页
Chinese Journal of Computers
基金
国家自然科学基金(61572123
61502092
61672148)
国家杰出青年基金(71325002)
中央高校基本科研业务费专项资金资助项目(N151604001)
中国博士后科学基金(2016M591449)
教育部-中国移动科研基金(MCM20160201)
辽宁省百千万人才工程项目(201514)资助~~
关键词
位置隐私保护
K-匿名
粒子群优化算法
隐马尔可夫模型
移动通信网络
location privacy protection
k-anonymity
particle swarm optimization
hidden Markov model
mobile communication networks
