一种基于区间验证的隐私保护真值发现算法  

Privacy-preserving Truth Discovery Algorithm Based on Interval Verification

作　　者：张志敏[1] 赵国生[1] 王健[2] ZHANG Zhimin;ZHAO Guosheng;WANG Jian(School of Computer Science and Information Engineering,Harbin Normal University,Harbin 150025,China;School of Computer Science and Technology,Harbin University of Science and Technology,Harbin 150080,China)

机构地区：[1]哈尔滨师范大学计算机科学与信息工程学院,哈尔滨150025 [2]哈尔滨理工大学计算机科学与技术学院,哈尔滨150080

出　　处：《小型微型计算机系统》2025年第2期410-416,共7页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61202458,61403109)资助;黑龙江省自然科学基金项目(LH2020F034)资助。

摘　　要：隐私保护真值发现技术在移动群智感知网络领域中受到了广泛关注.然而在实际应用中,恶意用户上传的异常值对真值发现结果的可靠性带来了较大影响.为此,提出了一种基于区间验证的隐私保护真值发现算法IVPPTD (Interval Verification based Privacy-Preserving Truth Discovery).首先,采用Paillier同态加密方法实现用户感知数据的安全上传和真值发现,保护用户的感知数据、权重信息以及估算真值的隐私不被泄露.其次,提出一种密文域中的异常数据过滤算法,对数据约束区间外的异常值进行数据清洗,从而在保护用户敏感信息不被泄露的前提下,提高真值发现结果的可靠性.最后,基于感知平台和密钥生成中心协作完成真值发现过程,减少了用户与云服务器之间的通信开销.仿真实验结果表明,所提方法具有高准确率、对异常值的鲁棒性以及较低的计算开销.Privacy-preserving truth discovery technology has received widespread attention in the field of mobile crowdsensing networks.However,in practical applications,the abnormal values uploaded by malicious users have a significant impact on the reliability of truth discovery results.Therefore,a privacy-preserving truth discovery algorithm based on interval verification is proposed.Firstly,the Paillier homomorphic encryption method is used to achieve secure crowdsensing data uploading and truth discovery,protecting the privacy of user crowdsensing data,weight information,and estimated truth from being compromised.Secondly,an exception value filtering algorithm is proposed to clean outliers outside the data constraint range,which can improve the reliability of truth discovery results while protecting sensitive user information from leakage.Finally,based on the collaboration between the crowdsensing platform and the key generation center,the truth discovery process is completed,reducing the communication overhead between users and cloud servers.The simulation experimental results show that the proposed method has high accuracy,robustness to outliers,and low computational overhead.

关 键 词：移动群智感知 真值发现 隐私保护 异常数据过滤 同态加密 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]